Endometrial Glandular And Stromal Breakdown Research Articles

The expression pattern of CD10 and CD31 identifies fine fibrovascular stroma of grade 1-endometrial endometrioid carcinomas in cytology.

The objective of this study was to assess the diagnostic utility of CD10 in the differential diagnosis of grade 1-endometrial endometrioid carcinoma (G1-EEC) and the metaplastic changes associated with the endometrial glandular and stromal breakdown (EGBD) on liquid-based cytological (LBC) samples. (1) The type and distribution of CD10-positive cells in EGBD and G1-EEC patients were evaluated. (2) Based on the results from (1), histological and cytological specimens were double-immunostained with CD31 and CD10 to confirm whether CD10-positive tubular-canalicular material found in (1) was represented by fine threads of endometrial-type fibrovascular stroma. (3) Based on the results from (2), additional immunostaining of histological specimens was performed for CD146 and αSMA as markers of perivascular cells. (1) CD10 positive cells showed two main patterns of expression: cytoplasmic immunoreactivity in the form of dense brown granules in EGBD and tubular-canalicular branching patterns in G1-EEC. (2) The tubular-canalicular material observed in cytological specimens of G1-EEC samples co-expressed CD10 and CD31, and was interpreted as representing fine threads of endometrial fibrovascular stroma in the corresponding histological samples. Conversely, metaplastic changes in EGBD cases, only a few CD31-positive signals were found inside the condensed stromal clusters with CD10-positive. (3) Cells surrounding the CD31-positive vascular endothelial cells expressed CD146 and αSMA; moreover, some of the thin CD10-positive fibrous stromal strands also co-expressed αSMA. CD10 is a very useful immunomarker for distinguishing between G1-EEC and the metaplastic changes of EGBD in LBC samples.

Insulin-like growth factor-II mRNA-binding protein 3 immunocytochemical expression in direct endometrial brushings: Possible diagnostic help in endometrial cytology.

This study evaluated the immunocytochemical (ICC) expression of IMP3 in direct endometrial brushings processed as liquid-based cytology (LBC) samples of endometrioid adenocarcinoma (EAC), serous carcinoma (ESC) and surface papillary syncytial change (SPSC) with endometrial glandular and stromal breakdown (EGBD) to exploit its possible differential diagnostic aid. In total, 333 samples of LBC samples were obtained from selected outpatients in parallel with Pipelle endometrial sampling. They consisted of 97 EAC (83 grade 1: EAC-1, 14 EAC-3), 35 ESC and 201 benign endometrial samples (51 proliferative, 42 secretory, 38 atrophic, 70 SPSC with EGBD). ICC expression of insulin-like growth factor-II mRNA-binding protein 3 (IMP3) was manually performed on Papanicolaou-stained LBC samples. The ESC samples showed positive staining cells in 100%, EAC-3 in 28.5%, and EAC-1 in 2.4% cases. All the benign endometrium samples were negative. Only ESC cases showed strong immunoreactivity (≥3+) in more than 50% of tumour cells with an average frequency of 80%. IMP3 is a helpful immunomarker to distinguish ESC from EAC and SPSC in endometrial cytology.

Inter-observer reproducibility of endometrial cytology by the Osaki Study Group method: utilising the Becton Dickinson SurePath™ liquid-based cytology.

The purpose of the present study was to evaluate the reproducibility of the cytological diagnosis of endometrial lesions by the Osaki Study Group (OSG) method of new cytological diagnostic criteria using BD SurePath™ (SP)-liquid-based cytology (LBC). This cytological classification using the OSG method consists of six categories: (i) normal endometrium (NE), (ii) endometrial glandular and stromal breakdown (EGBD), (iii) atypical endometrial cells, cannot exclude atypical endometrial hyperplasia or more (ATEC-A), (iv) adenocarcinoma including atypical endometrial hyperplasia or malignant tumour (Malignancy), (v) endometrial hyperplasia without atypia (EH) and (vi) atypical endometrial cells of undetermined significance (ATEC-US). For this study, a total 244 endometrial samplings were classified by two academic cytopathologists as follows: 147 NE cases , 36 EGBD cases , 47 Malignant cases, eight ATEC-A cases, two EH cases and four ATEC-US cases. To confirm the reproducibility of the diagnosis and to study the inter- and intra-observer agreement further, a second review round followed at 3-month intervals, which included three additional cytopathologists. The inter-observer agreement of NE classes improved progressively from 'good to fair' to 'excellent', with values increasing from 0.70 to 0.81. Both EGBD and Malignancy classes improved progressively from 'good to fair' to 'excellent', with values increasing from 0.62-0.63 to 0.84-0.95, respectively. The overall intra-observer agreement between the first and the second rounds was 'good to fair' to 'excellent', with values changing from 0.79 to 0.85. All kappa improvements were significant (P < 0.0001). In this study, it seemed that the use of the OSG method as the new diagnostic criteria for SP-LBC preparation, may be a valid method to improve the precision (reproducibility) of endometrial cytology.

Diagnostic Utility of Notch-1 Immunocytochemistry in Endometrial Cytology

Objective: It was the aim of this study to evaluate the diagnostic utility of Notch-1 immunocytochemistry in distinguishing endometrial glandular and stromal breakdown (EGBD) from endometrial adenocarcinoma in endometrial cytology. Study Design: Samples of normal endometrium, EGBD and endometrial adenocarcinoma were subjected to immunocytochemical staining for Notch-1, and we examined the labeling index (LI) of Notch-1 (the ratio of intranuclear Notch-1-positive cells to total cells). We compared (1) the Notch-1 LI in normal endometrium, (2) the Notch-1 LI between normal endometrium and endometrial adenocarcinoma, and (3) the Notch-1 LI in normal endometrium, EGBD and endometrial adenocarcinoma. Results: In analysis item 1, the LI of Notch-1 was 32.9 ± 8.4, 19.4 ± 8.2 and 12.5 ± 7.5% in proliferative endometrium, secretory endometrium and atrophic endometrium, respectively. In analysis item 2, the LI of Notch-1 in endometrial adenocarcinoma was 45.2 ± 7.4%, which was significantly higher than that in normal endometrium. In analysis item 3, the LI of Notch-1 in EGBD was 31.3 ± 8.3%, which was significantly lower than that in endometrial adenocarcinoma. Conclusion: In conclusion, Notch-1 immunocytochemistry is a useful method for distinguishing between EGBD and endometrial carcinoma in endometrial cytology.

Nuclear features in endometrial cytology: Comparison of endometrial glandular and stromal breakdown and endometrioid adenocarcinoma grade 1

This study was to clarify the nuclear features of "condensed clusters of stromal cells (EGBD-stromal cells)" and "metaplastic clumps with irregular protrusions (EGBD-metaplastic cells)" which may be recognized in endometrial glandular and stromal breakdown (EGBD) cases in liquid-based cytologic (LBC) preparations of endometrial brushings. The material consists of cytologic smears of 20 cases of proliferative endometrium (PE), 20 cases of EGBD, and 20 cases of endometrioid adenocarcinoma grade 1 (G1) for which histopathological diagnosis was obtained by endometrial curettage. Nuclear findings were examined in PE cells, EGBD-stromal cells, EGBD-metaplastic cells, and G1 cells, respectively. It was examined about the following items: (1) Nuclear shape; (2) A long/minor axis ratio in cell nuclei; (3) An area of cell nuclei; (4) Overlapping nuclei; (5) The distribution pattern of nuclei within cell clusters. The following observations were made: (1) In PE cells, round-oval nuclei appeared to predominate, overlapping nuclei were not observed, and a slightly abnormal distribution pattern of nuclei was recorded; (2) In EGBD-stromal cells, reniform nuclei were characteristically observed, nuclei had small size and a generally elongated appearance, overlapping nuclei were recognized, and a remarkable abnormal distribution pattern of nuclei was found; (3) In EGBD-metaplastic cells, spindle nuclei were a characteristic feature, nuclei were larger in size and had a bipolar appearance, overlapping nuclei with moderately abnormal distribution pattern of nuclei were identified; (4) In G1 cells round-oval nuclei predominated, overlapping nuclei with moderately abnormal distribution pattern of nuclei were found. The study demonstrates that the analysis of selected nuclear findings appears to be very useful in the cytopathological assessment of endometrial lesions in LBC samples, especially for the discrimination of EGBD and G1.

Expression of immunoreactivity of nuclear findings by p53 and cyclin a in endometrial cytology: Comparison with endometrial glandular and stromal breakdown and endometrioid adenocarcinoma grade 1

It is well known that "condensed cluster of stromal cells (CCSC)" and "metaplastic clumps with irregular protrusion (MCIP)" in endometrial glandular and stromal breakdown (EGBD) cases may simulate "clumps of cancer cells (CCC)" in endometrioid adenocarcinoma grade 1 (G1), leading to difficulty in cytological interpretation. The aim of this study was undertaken to clarify the cytological immunoreactivity of nuclear findings about CCSC and MCIP which may be recognized in EGBD cases by using p53 protein and cyclin A in liquid-based cytologic (LBC) preparations. The material consists of cytologic smears of 20 cases of EGBD and 20 cases of G1 for which histopathological diagnosis was obtained by endometrial curettage at the JA Suzuka General Hospital. The evaluation of immunoreactivity was performed by using the intensity of nuclear staining and the nuclear labeling index (N-LI). The intensity of nuclear staining was scored as negative (0), weak (1), moderate (2), or strong (3). The N-LI was scored as less than 10% (0), from 10 to 25% (1), from 26 to 50% (2), or greater than 50% (3). The final score was calculated of the addition of both partial scores. Results are as follows: As for the p53 protein immunoreactivity, CCC (2.4 ± 1.4) was a significantly higher value in comparison with CCSC (0) and MCIP (0.8 ± 0.4), respectively. As for the cyclin A immunoreactivity, CCC (2.8 ± 1.1) was a significantly higher value in comparison with CCSC (0) and MCIP (0.6 ± 0.5), respectively. CCSC and MCIP in EGBD are misunderstood as cellular atypia and structural atypia on occasion; but, as for results of the immunoreactivity scores of p53 protein and cyclin A in our study, it seemed that those biochemical characters proved that the biological activity level was low (or degenerative). The results of the current study demonstrated that the cytological immunoreactivity of nuclear findings by p53 and cyclin A appear to be more useful for the LBC assessment of endometrial lesions, especially for the discrimination of EGBD and G1.

Nuclear characteristics of the endometrial cytology: Liquid‐based versus conventional preparation

The aim of this study was to assess the utility of liquid-based cytologic preparation (LP) compared with conventional preparation (CP) for the assessment of nuclear findings in endometrial glandular and stromal breakdown (EGBD) which may be misdiagnosed as carcinoma in EGBD cases. The material consists of cytologic smears including 20 cases of proliferative endometrium (PE), 20 cases of EGBD, and 20 cases of endometrioid adenocarcinoma grade1 (G1) for which histopathological diagnosis was obtained by endometrial curettage at the JA Suzuka General Hospital. Nuclear findings were examined in PE cells, EGBD-stromal cells, EGBD-metaplastic cells, and G1 cells, respectively. It was examined about the following items; (1) nuclear shape; (2) A long/minor axis ratio in cell nuclei; (3) an area of cell nuclei; (4) overlapping nuclei. Results are as follows: (1) nuclear shape; as for the reniform shape of EGBD-stromal cells and spindle shape of EGBD-metaplastic cells, the ratio of the LP method was a higher value than the CP method. (2) The long axis and area of cell nuclei; LP in all groups was a recognizable tendency for nuclear shrinkage. (3) The long/minor axis ratio in cell nuclei; only EGBD-metaplastic cells recognize a significant difference between CP and LP. (4) Overlapping nuclei; LP was a higher value in comparison with CP in the other groups except PE cells, and the degree of overlapping nuclei was enhanced about three times. Therefore, although a cell of LP has a shrinking tendency, (1) it is excellent that LP preserves a characteristic of nuclear shape than CP; (2) a cellular characteristic becomes clearer, because three-dimensional architecture of LP is preserved of than CP. As for the standard preparation method for endometrial cytology samples, we considered that a concrete introduction of the LP method poses no problems.

Endometrial glandular and stromal breakdown, part 4: Cytomorphology of “condensed cluster of stromal cells including a light green body”

This study was undertaken to clarify the origin of the chromophilic substance that stained in light green [light green body (LGB)] was observed in the condensed cluster of stromal cells and in the background of endometrial glandular and stromal breakdown cases. The material consists of cytologic smears of endometrial glandular and stromal breakdown (EGBD) from 58 samples in which a histopathological diagnosis was subsequently obtained by endometrial curettage. The following parameters were examined. (1) The occurrence and number of a LGB. As for the "condensed cluster of stromal cells," "condensed cluster of stromal cells including LGB" and "LBG in background," it was observed with all fields on one slide per case. When it was observed more than one in a preparation, the occurrence was determined and the number was also calculated. (2) Immunostaining of paraffin-embedded tissue sections and cytologic preparations. 1 The occurrence of "condensed cluster of stromal cells," "condensed cluster of stromal cells including LGB" and "LGB in background" was 100%, 44.8% and 91.4%, respectively. The number of "condensed cluster of stromal cells," "condensed cluster of stromal cells including LGB" and "LGB in background" was 21.2, 2.0 and 4.0, respectively. 2 When histological sections from paraffin-embedded tissue were tested by immunocytochemistry, "LGB" stained weakly or negative for fibrinogen, whereas it stained positively for CD31 and factor VIII. In addition, it stained positively for CD42b. 3 When cytologic preparation was tested by immunocytochemistry, "LGB" stained positively for CD31, factor VIII and CD42b. However, as for the fibrinogen, quantitative and qualitative evaluation of the immunocytochemical stain was not feasible due to extensive nonspecific reaction in the whole preparation. The "LGB" in EGBD cases proved to be a thrombus mostly made up by platelets, and it seemed that the recognition of the LGB showed to be a useful cytomorphological criterion for an always more accurate diagnosis of EGBD cases.

Endometrial glandular and stromal breakdown, Part 3: Cytomorphology of “condensed cluster of stromal cells”

The aim of this study was undertaken to clarify the cytological characteristic of the "condensed clusters of stromal cells," which may be recognized in endometrial glandular and stromal breakdown (EGBD) cases. The material consists of 60 cases of cytologic smears for which histopathological diagnosis was obtained by endometrial curettage; they comprised 30 cases of EGBD and 30 cases of endometrioid adenocarcinoma grade 1 (G1). The following parameters were examined for "condensed clusters of stromal cells" in EGBD and for "clumps of cancer cells" in G1, respectively: (1) the occurrence of "condensed clusters of stromal cells"; (2) the nuclear shape; (3) a longer/shorter axis ratio in cell nuclei; (4) the area of cell nuclei; (5) the presence of overlapping nuclei; (6) nuclear crowding; (7) immunostaining. (1)"Condensed clusters of stromal cells" were only observed in EGBD. (2) A reniform nuclear shape was observed in 100% EGBD (P < 0.0001) in comparison to G1 (3%). (3) The longer/shorter axis ratio in cell nuclei, G1 (1.37 +/- 0.2) was significantly lower in comparison to EGBD (1.53 +/- 0.12, P = 0.0005). (4) Nuclear area in G1 (51.6 +/- 11.9, P < 0.0001) was significantly higher in comparison to EGBD (24.3 +/- 3.9 microm(2)). (5) The score of overlapping nuclei in EGBD (2.5 +/- 0.49) was significantly higher in comparison to G1 (1.8 +/- 0.44, P < 0.0001). (6) The nuclear crowding score was the same both in EGBD (2) and G1 (2) and these findings were not statistically significant. (7) Both CD10 and Wilms' tumor protein 1 were positive in the "condensed clusters of stromal cells" in the EGBD. The anti-cytokeratin staining was positive in "clumps of cancer cells" in the G1. The evaluation of the immunocytochemical findings by combining the Wilms' tumor 1 protein, CD10, and the anti-cytokeratin with the considered cytomorphologic features (reniform nucleus) may be useful for a correct diagnosis of EGBD in endometrial cytology.

Expression of immunoreactivity and genetic mutation in eosinophilic and ciliated metaplastic changes of endometrial glandular and stromal breakdown: cytodiagnostic implications

Various metaplastic changes may be present in endometrium, in which also cellular atypias may often be observed. Particularly, eosinophilic and ciliated changes (ECCs) occur in both nonneoplastic and neoplastic endometrium. This may cause confusion in the cytodiagnosis. This study was enterprised to investigate the possible help of immunocytochemical and cytogenetic study in the diagnostic and biologic assessment of ECC cells. In immunocytochemistry for p53 protein, Ki-67, and cyclin A, the material consists of 40 cases of cytologic smears examined by direct sampling of the endometrial cavity comprising 30 cases of ECC in endometrial glandular and stromal breakdown (EGBD) and 10 cases of well-differentiated adenocarcinoma (G1). After cytodiagnosis, immunostaining for p53 protein, Ki-67, and cyclin A was performed on multiple wet-fixed slides from each single case to evaluate the immunoreactivity, intensity of nuclear staining, and nuclear labeling index (N-LI). The intensity of nuclear staining was scored as negative (0), weak (1), moderate (2), or strong (3), and the N-LI was scored as less than 10% (0), from 10% to 25% (1), from 26% to 50% (2), or more than 50% (3), and the final score was calculated by adding both partial scores. A statistical significance test was performed by using Mann-Whitney U test, and the result was judged as significant when the P value was less than .05. For genetic mutation analysis of p53, the material comprised 6 cases of EGBD in which a high score was measured with immunocytochemistry for p53 protein, and the presence of ECC was confirmed on the hematoxylin and eosin. The ECC cells on paraffin-embedded specimens were captured using laser capture microdissection technology. Mutations in p53 gene (exons 5-8) were examined using DNA sequencing analysis. In immunocytochemistry for p53 protein, Ki-67, and cyclin A, the proportions of immunoreactive cells for p53 were statistically higher in ECC compared with those of G1 (P < .05). The proportions of the immunoreactive cells for Ki-67 and cyclin A were statistically higher in G1 compared with those of ECC (P < .05). (2) In genetic mutation analysis of p53, DNA sequencing of p53 in 6 cases revealed no mutations. The percentage of immunoreactive cells for p53 protein were higher in ECC than in G1, but the mutation point was not confirmed in genetic mutation analysis. The differential expression of these biologic parameters in ECC cells could be considered of possible relevance to the cytopathologic diagnosis in the future.

Utility of thin‐layer preparations in endometrial cytology: Immunocytochemical expression of PTEN, beta‐catenin and p53 for benign endometrial lesions

This article focuses on the characteristic features of morphology and molecular biology of PTEN, beta-catenin, and p53 immunocytochemistry in normal endometrium (proliferative, secretory, and atrophic) and endometrial glandular and stromal breakdown (EGBD) using thin-layer specimens. During a 6-month period, 120 endometrial samples were collected directly using the Uterobrush and a thin-layer specimen was prepared. Immunocytochemical expression of PTEN, beta-catenin, and p53 were investigated using 30 cases each of proliferative endometrium (PE), secretory endometrium (SE), atrophic endometrium (AE), and EGBD.PTEN expression of normal endometrial glandular epithelial cells changes with the hormonal status; PE produce very high expression, SE creates attenuation or disappearance of PTEN expression and AE diminished more in comparison with SE. PTEN expression of EGBD showed a tendency towards attenuation compared with PE, but showed high expression in comparison with SE and AE. As for the immunoreactivity of beta-catenin, in all phases (PE, SE, AE, and EGBD), it was observed in the cytoplasm of glandular epithelial cells, but not nuclei, and showed strong membranous staining. As for the p53 immunoreactivity, p53 positivity was not observed in the glandular epithelial cells in all phases (PE, SE, AE, and EGBD) with the exception of some metaplastic cells. The presence of p53 immunoreactivity of a weak, low ratio in metaplastic cells was unexpected. In the current study, the expression manner of PTEN, beta-catenin, and p53 immunocytochemistry was observed in the normal endometrium (PE, SE, and AE) and EGBD.

Utility of thin-layer preparations in the endometrial cytology

The purpose of the current study was to examine the use of thin-layer cytologic (TLC) preparation compared to conventional cytologic preparation (CCP) in the normal endometrium (proliferative, secretory, atrophic) and endometrial glandular and stromal breakdown (EGBD). During a 6-month period, we compiled 158 cases by collecting a direct endometrial sample using the Uterobrush. The material comprised 40 cases of proliferative endometrium, 42 cases of secretory endometrium, 46 cases of atrophic endometrium, and 30 cases of EGBD. The following points were investigated: (1) number of endometrial epithelial cell clumps; (2) presence of TLC > CCP cases on number of epithelial cell clumps; (3) number of condensed cluster of stromal cells; (4) presence of TLC > CCP cases on number of condensed cluster of stromal cells; (5) presence of metaplastic clumps with irregular protrusion-containing condensed stromal cluster; (6) presence of a clear background; (7) presence of blood vessel in TLC; (8) presence of blood vessel of length more than diameter of a field in object x20 glasses in TLC. (1) In all phases, the number of epithelial cell clumps per a unit area of a preparation of TLC is greater than in CCP. (2) Cells (condensed cluster of stromal cells and metaplastic clumps with irregular protrusion-containing condensed stromal cluster) of useful and adequate numbers for a diagnosis of EGBD were observed in TLC. (3) In all phases, TLC was significantly higher than CCP on the appearance of a clear background. (4) The proliferative endometrium and secretory endometrium were highly significant in comparison with atrophic endometrium and EGBD, respectively, in terms of the occurrence of a blood vessel of length more than diameter of a field in object x20 glasses. Although the preparation area of TLC is smaller than that of CCP, the preparation has a clean background so that an accurate report on the patient's condition is possible. Therefore, TLC preparation is a useful tool for the accurate and reliable diagnosis of normal endometrial phase and EGBD, because the preparation area is confined and identification of the target cell clumps is easy.

Endometrial glandular and stromal breakdown, part 1: Cytomorphological appearance

Endometrial carcinoma is the most common invasive neoplasm of the female reproductive tract. Early detection and accurate diagnosis of these lesions and its precursor by endometrial cytology is now accepted in Japan and regarded as an effective primary method of evaluating endometrial pathology (atypical hyperplasia or carcinoma). Careful cytomorphologic evaluation of the abnormal endometrial lesions has made possible an accurate and reproducible microscopic assessment. The current study was conducted to determine the significance of endometrial cytology on disordered endometrium associated with anovulation when compared with endometrial hyperplasia. From January 1998 through April 2004, 144 cases on which histopathological diagnoses were obtained by endometrial curettage after taken direct endometrial sample by Endocyte. The materials comprise 49 cases of normal proliferative endometrium, and 63 cases of endometrial hyperplasia without atypia were prepared as control cases. The cytomorphology was examined involving so-called endometrial glandular and stromal breakdown (EGBD). EGBD cases evidenced significant numbers of stromal cells condensed and formed compact nests with hyperchromatic nuclei and little or no cytoplasm. They were often associated with fragmented clusters of endometrial glands with condensed cluster of stromal cells. Both the fragmented cluster of endometrial glands and condensed cluster of stromal cells are a characteristic cytologic feature of EGBD endometrium on the cyto-architectural diagnosis. The combination of these cellular patterns is highly specific to this abnormal pathological condition in EGBD endometrium. To improve the accuracy of the cytodiagnosis, it is important that the cytology of the EGBD endometrium should be diagnosed negative; as a result, we can achieve successful endometrial cytology with cyto-architectural criteria for the endometrial pathology.

Endometrial glandular and stromal breakdown, part 2: Cytomorphology of papillary metaplastic changes

Careful cytomorphologic evaluation of abnormal endometrial lesions has made accurate and reproducible microscopic assessment possible. Histopathology of patients with dysfunctional uterine bleeding due to an anovulatory cycle usually contain endometrial glandular and stromal breakdown (EGBD) and papillary metaplasia on the endometrial surface epithelium, if an appropriate sample has been collected. We often recognized abnormal cell clumps in the cytological smears with EGBD cases. They were composed of metaplastic cells, and some irregular small projection figures were observed from the margins of the cell clumps. We describe the quantitation of metaplastic clumps with irregular protrusions (MCIP) in endometrial endocyte samples. The current study presents the cytomorphological characteristics of the metaplastic changes recognized in EGBD cases. During a 7-yr period, 144 cases for which histopathological diagnoses were obtained following endometrial curettage, after collecting a direct endometrial sample using the endocyte. The material comprised 49 cases of normal proliferative endometrium (NPE) (patients aged 28-51, average 39.9), 32 cases of EGBD (patients aged 30-67, average 49.6), and 63 cases of endometrial hyperplasia without atypia (EH) (patients aged 35-65, average 47.7). The following points were investigated: (1) the occurrence of metaplastic cells; (2) the occurrence and the frequency of MCIP; and (3) the occurrence of MCIP that contains condensed stromal clusters. Metaplastic cells were seen in 93.8% of the EGBD cases. Cytomorphologic pattern identified with MCIP was 90.6%, and its frequency showed 16.1%. The occurrence of MCIP that contain condensed stromal clusters (93.1%) showed a strong association in comparison with other lesions, such as NPE and EH. Our data appear to indicate that the appearance of MCIP with condensed stromal clusters originated from the papillary metaplasia, which occurred on the endometrial surface epithelium. The cytologic observation of those cells may be a useful indicator for providing the nature of EGBD endometrium.