Performance of hysteroscopy in diagnosing chronic endometritis and the role of intra- and inter-observer variability: a prospective study of 70 cases

Diagnostic HSC was scheduled during the late follicular phase of the cycle (between days 8 and 12). The examination steps were as follows: – Introduction of a 2.9 mm 30° rigid hysteroscope (KARL STORZ) with distension of the uterine cavity by pressurized infusion of physiological saline (0.9% NaCl). – Examination of the uterine cavity and the cervico-isthmic canal using a “no touch” technique to avoid touching the uterine mucosa and thereby ensure accurate video interpretation. – Endometrial biopsy performed at the end of the examination using a 23 cm NOVACK curette. Introduction of a 2.9 mm 30° rigid hysteroscope (KARL STORZ) with distension of the uterine cavity by pressurized infusion of physiological saline (0.9% NaCl). Examination of the uterine cavity and the cervico-isthmic canal using a “no touch” technique to avoid touching the uterine mucosa and thereby ensure accurate video interpretation. Endometrial biopsy performed at the end of the examination using a 23 cm NOVACK curette. The HSC videos were recorded and presented blind to two specialized reproductive medicine observers. A second interpretation of these videos was requested from the same observers after 4 weeks. The observers were required to answer the following three questions: 1. Is the video interpretable? 2. Does the hysteroscopic appearance suggest CE? 3. If yes, which hysteroscopic features suggesting the diagnosis of CE according to Cicinelli et al are observed: 1. Micropolypes ( Figure 1 ). Figure 1 (A) Diffuse micro-polyps observed via hysteroscopy in liquid medium in a patient with CE. (B) Sporadic micro-polyps observed via hysteroscopy in liquid medium in a patient with CE. (C and D) Few sporadic micro-polyps observed via hysteroscopy in liquid medium in a patient without CE Figure 1 Halouani. Performance of hysteroscopy in diagnosing chronic endometritis and the role of intra- and inter-observer variability. AJOG Glob Rep 2025. 2. Hemorrhagic spots ( Figure 2, A and B ). Figure 2 (A) Hemorrhagic spots observed via diagnostic hysteroscopy in liquid medium in a patient without CE. (B) Hemorrhagic spots on the endometrium in a patient with CE. (C and D) Strawberry-like appearance observed via diagnostic hysteroscopy in liquid medium in two patients without CE Figure 2 Halouani. Performance of hysteroscopy in diagnosing chronic endometritis and the role of intra- and inter-observer variability. AJOG Glob Rep 2025. 3. Strawberry aspect ( Figure 2, C and D ). 4. Stromal edema ( Figure 3, A and B ). Figure 3 (A) Stromal edema and few micro-polyps observed via diagnostic hysteroscopy in the late follicular phase in a patient without CE. (B) Stromal edema observed via diagnostic hysteroscopy in liquid medium in the late follicular phase in a patient with CE. (C) Focal endometrial hyperemia observed via hysteroscopy in liquid medium in a patient with CE Figure 3 Halouani. Performance of hysteroscopy in diagnosing chronic endometritis and the role of intra- and inter-observer variability. AJOG Glob Rep 2025. 5. Focal hyperemia ( Figure 3, C ). 6. Diffuse hyperemia ( Figure 4, A and B ). Figure 4 (A) Diffuse endometrial hyperemia observed via diagnostic hysteroscopy in liquid medium in a patient with CE. (B) Diffuse endometrial hyperemia observed via diagnostic hysteroscopy in liquid medium in a patient without CE Figure 4 Halouani. Performance of hysteroscopy in diagnosing chronic endometritis and the role of intra- and inter-observer variability. AJOG Glob Rep 2025. Is the video interpretable? Does the hysteroscopic appearance suggest CE? If yes, which hysteroscopic features suggesting the diagnosis of CE according to Cicinelli et al are observed: 1. Micropolypes ( Figure 1 ). Figure 1 (A) Diffuse micro-polyps observed via hysteroscopy in liquid medium in a patient with CE. (B) Sporadic micro-polyps observed via hysteroscopy in liquid medium in a patient with CE. (C and D) Few sporadic micro-polyps observed via hysteroscopy in liquid medium in a patient without CE Figure 1 Halouani. Performance of hysteroscopy in diagnosing chronic endometritis and the role of intra- and inter-observer variability. AJOG Glob Rep 2025. 2. Hemorrhagic spots ( Figure 2, A and B ). Figure 2 (A) Hemorrhagic spots observed via diagnostic hysteroscopy in liquid medium in a patient without CE. (B) Hemorrhagic spots on the endometrium in a patient with CE. (C and D) Strawberry-like appearance observed via diagnostic hysteroscopy in liquid medium in two patients without CE Figure 2 Halouani. Performance of hysteroscopy in diagnosing chronic endometritis and the role of intra- and inter-observer variability. AJOG Glob Rep 2025. 3. Strawberry aspect ( Figure 2, C and D ). 4. Stromal edema ( Figure 3, A and B ). Figure 3 (A) Stromal edema and few micro-polyps observed via diagnostic hysteroscopy in the late follicular phase in a patient without CE. (B) Stromal edema observed via diagnostic hysteroscopy in liquid medium in the late follicular phase in a patient with CE. (C) Focal endometrial hyperemia observed via hysteroscopy in liquid medium in a patient with CE Figure 3 Halouani. Performance of hysteroscopy in diagnosing chronic endometritis and the role of intra- and inter-observer variability. AJOG Glob Rep 2025. 5. Focal hyperemia ( Figure 3, C ). 6. Diffuse hyperemia ( Figure 4, A and B ). Figure 4 (A) Diffuse endometrial hyperemia observed via diagnostic hysteroscopy in liquid medium in a patient with CE. (B) Diffuse endometrial hyperemia observed via diagnostic hysteroscopy in liquid medium in a patient without CE Figure 4 Halouani. Performance of hysteroscopy in diagnosing chronic endometritis and the role of intra- and inter-observer variability. AJOG Glob Rep 2025. Micropolypes ( Figure 1 ). Figure 1 (A) Diffuse micro-polyps observed via hysteroscopy in liquid medium in a patient with CE. (B) Sporadic micro-polyps observed via hysteroscopy in liquid medium in a patient with CE. (C and D) Few sporadic micro-polyps observed via hysteroscopy in liquid medium in a patient without CE Figure 1 Halouani. Performance of hysteroscopy in diagnosing chronic endometritis and the role of intra- and inter-observer variability. AJOG Glob Rep 2025. (A) Diffuse micro-polyps observed via hysteroscopy in liquid medium in a patient with CE. (B) Sporadic micro-polyps observed via hysteroscopy in liquid medium in a patient with CE. (C and D) Few sporadic micro-polyps observed via hysteroscopy in liquid medium in a patient without CE Hemorrhagic spots ( Figure 2, A and B ). Figure 2 (A) Hemorrhagic spots observed via diagnostic hysteroscopy in liquid medium in a patient without CE. (B) Hemorrhagic spots on the endometrium in a patient with CE. (C and D) Strawberry-like appearance observed via diagnostic hysteroscopy in liquid medium in two patients without CE Figure 2 Halouani. Performance of hysteroscopy in diagnosing chronic endometritis and the role of intra- and inter-observer variability. AJOG Glob Rep 2025. (A) Hemorrhagic spots observed via diagnostic hysteroscopy in liquid medium in a patient without CE. (B) Hemorrhagic spots on the endometrium in a patient with CE. (C and D) Strawberry-like appearance observed via diagnostic hysteroscopy in liquid medium in two patients without CE Strawberry aspect ( Figure 2, C and D ). Stromal edema ( Figure 3, A and B ). Figure 3 (A) Stromal edema and few micro-polyps observed via diagnostic hysteroscopy in the late follicular phase in a patient without CE. (B) Stromal edema observed via diagnostic hysteroscopy in liquid medium in the late follicular phase in a patient with CE. (C) Focal endometrial hyperemia observed via hysteroscopy in liquid medium in a patient with CE Figure 3 Halouani. Performance of hysteroscopy in diagnosing chronic endometritis and the role of intra- and inter-observer variability. AJOG Glob Rep 2025. (A) Stromal edema and few micro-polyps observed via diagnostic hysteroscopy in the late follicular phase in a patient without CE. (B) Stromal edema observed via diagnostic hysteroscopy in liquid medium in the late follicular phase in a patient with CE. (C) Focal endometrial hyperemia observed via hysteroscopy in liquid medium in a patient with CE Focal hyperemia ( Figure 3, C ). Diffuse hyperemia ( Figure 4, A and B ). Figure 4 (A) Diffuse endometrial hyperemia observed via diagnostic hysteroscopy in liquid medium in a patient with CE. (B) Diffuse endometrial hyperemia observed via diagnostic hysteroscopy in liquid medium in a patient without CE Figure 4 Halouani. Performance of hysteroscopy in diagnosing chronic endometritis and the role of intra- and inter-observer variability. AJOG Glob Rep 2025. (A) Diffuse endometrial hyperemia observed via diagnostic hysteroscopy in liquid medium in a patient with CE. (B) Diffuse endometrial hyperemia observed via diagnostic hysteroscopy in liquid medium in a patient without CE An IHC study of the expression of the plasma cell marker CD138 was systematically performed for all samples studied. We used the anti-CD138 monoclonal antibody (Clone Ml15) from Leica Biosystems (Reference: PA0088), ready to use on the BOND immunohistochemistry automated system. In cases of identifiable plasma cell infiltrate on histological examination and/or after IHC study, the number of plasma cells per high-power field (HPF) (×400) was specified. We used a threshold of at least five plasma cells per HPF (×400) to diagnose CE ( Figure 5 ). Figure 5 Chronic endometritis: Plasma cells in the stromal chorion observed with IHC CD-138. (1) Low magnification and (2) high magnification (×400) Figure 5 Halouani. Performance of hysteroscopy in diagnosing chronic endometritis and the role of intra- and inter-observer variability. AJOG Glob Rep 2025. Chronic endometritis: Plasma cells in the stromal chorion observed with IHC CD-138. (1) Low magnification and (2) high magnification (×400)

Ahmed Halouani: Formal analysis. Haithem Aloui: Writing – original draft, Conceptualization. Rim Hamdaoui: Software. Yassine Masmoudi: Investigation. Amel Triki: Validation, Supervision. Anissa Ben Amor: Validation. Lazhar Halouani: Supervision.

This is a prospective, multicenter observational study conducted from June 6, 2021, to August 8, 2022, at University Hospital Mongi Slim La Marsa and the ART Center Les Jasmins, Tunis, Tunisia. The study included women aged 18 to 42 years, consulting for infertility, with a pelvic ultrasound showing no uterine abnormalities. Excluded from the study were patients with biological inflammatory syndrome, those on antibiotic therapy or with a recent acute genital infection (less than 3 months), autolyzed biopsies, unusable HSC videos, atypical hyperplasia, or endometrial cancer. We calculated the sample size based on the prevalence of CE in the literature among infertile women, which ranges from 2.8% to 30%. 9 We considered a significance level ( P ≤.05), a 95% confidence level, and the values α =0.05 and i =0.02, using the following formula 10 : N = z a * P * ( 1 − P ) ( i 2 ) N : sample size Zα : confidence level. For a 95% confidence level, Zα =1.96 P : the average prevalence of CE is 16.4% Sample size N =67.18 (68 women).

A total of 70 patients were definitively included in this study based on the inclusion and exclusion criteria. The mean age of the cohort was 34.3±4.6 years. The average body mass index (BMI) was 28.9±5.3 kg/m². In this population, 95.7% of the patients were non-smokers. The mean duration of infertility was 60±15 months. Within our cohort, 79% of the patients presented with primary infertility, while 21% presented with secondary infertility. This difference was not statistically significant ( P =.80). We observed a significant association between younger age and the risk of CE ( P <.0001). Furthermore, there was a statistically significant correlation between increasing BMI and elevated risk of CE ( P <.0001). There was no significant association between smoking status and CE (96.7% vs 95%, p=1). The prevalence of CE confirmed by histopathological examination was 42.9%. The diagnostic performance of HSC for CE was assessed in comparison with histopathological findings. The sensitivity of HSC for detecting CE was 47.5%, with a specificity of 64.38%. The positive predictive value (PPV) of HSC for diagnosing CE was 50%, and the negative predictive value (NPV) was 62.05%. A statistically significant association was found between the presence of diffuse hyperemia observed during hysteroscopic examination and the risk of CE ( P =.0003). However, other hysteroscopic findings did not show a significant correlation with the risk of CE ( P >.05) ( Table 1 ). Table 1 Frequency of different hysteroscopic aspects of CE Table 1 Hysteroscopic signs Total With CE Without CE P Micro-polyps 8.57% (6/70) 3.33% 10% .7 Hemorrhagic spots 41.42% (29/70) 50% 35% .28 Focal hyperemia 77.14% (54/70) 86.7% 70% .31 Diffuse hyperemia 44.28% (31/70) 73.3% 22.5% .0003 Strawberry appearance 34.28% (24/70) 36.7% 32.5% .75 Stromal edema 74.28% (51/70) 90% 62.5% .095 Boldface indicates statistical significance (p<0.05). Halouani. Performance of hysteroscopy in diagnosing chronic endometritis and the role of intra- and inter-observer variability. AJOG Glob Rep 2025. Frequency of different hysteroscopic aspects of CE Boldface indicates statistical significance (p<0.05). Inter-rater reliability analysis indicated a strong agreement between the two observers in the hysteroscopic diagnosis of CE across the first and second readings, with Cohen’s kappa coefficient ( κ ) =0.62 ( Table 2 ). Intra-rater reliability analysis demonstrated moderate agreement, with Cohen’s kappa coefficient ( κ )=0.58 ( Table 3 ). Table 2 Inter-observer variability Table 2 Observer 1 Observer 2 Cohen’s kappa Inter-observer agreement 1st reading 1st reading 0.70 Strong agreement 2nd reading 2nd reading 0.50 Moderate agreement 1st reading 2nd reading 0.70 Strong agreement 2nd reading 1st reading 0.44 Moderate agreement Total 0.62 Strong agreement Halouani. Performance of hysteroscopy in diagnosing chronic endometritis and the role of intra- and inter-observer variability. AJOG Glob Rep 2025. Table 3 Intra-observer variability Table 3 The observers Cohen’s kappa Intraobserver agreement Observer 1 0.5 Moderate agreement Observer 2 0.68 Strong agreement Total 0.58 Moderate agreement Halouani. Performance of hysteroscopy in diagnosing chronic endometritis and the role of intra- and inter-observer variability. AJOG Glob Rep 2025. Inter-observer variability Intra-observer variability

HSC does not appear to be the diagnostic method of choice for CE. Endometrial biopsy using the Novak curette, combined with CD138 IHC study, proves to be more effective and less costly and can be performed on an outpatient basis. Larger-scale studies are needed to confirm these findings.

The prevalence of CE in this study was 42.9%, which was higher than the range reported in the literature among infertile women (2.8%–30%). 9 This could have indicated a selection bias. In fact, HSC was not routinely performed in our clinical practice; it was usually reserved for women with abnormal ultrasound findings or a history of previous ART failure. As women with abnormal ultrasound were excluded from the study, it was likely that our sample included a higher proportion of patients with recurrent implantation failure, a condition in which CE has been more frequently observed. This may have helped explain the higher prevalence of CE found in our population. The diagnostic accuracy of HSC for detecting CE in our study was moderate, with a sensitivity of 47.5%, specificity of 64.38%, PPV of 50%, and NPV of 62.05%. The level of agreement between different observers (inter-observer reliability) in diagnosing CE using HSC was strong, with a Cohen’s kappa ( κ ) of 0.62. However, the consistency of diagnoses made by the same observer over time (intra-observer reliability) was moderate, with a Cohen’s kappa ( κ ) of 0.58. The physicians participating in this study possessed substantial expertise and experience in their specialty. The HSC evaluations were presented to the two observers as videos rather than static images, facilitating a more comprehensive analysis of the uterine cavity. Follow-up data for patients diagnosed with CE, such as histological control, pregnancy, and birth rates post-treatment, were not reported, as these were not among the study’s objectives. A prospective study published in 2020 by Liu et al, 11 which included 320 women monitored for infertility or recurrent miscarriage, found a statistically significant association between the presence of diffuse or focal hyperemia, hemorrhagic spots, micropolypes, polyps, and CE. The combination of more than two hysteroscopic criteria could improve the specificity of HSC in diagnosing CE up to 99%. 12 The sensitivity of HSC in diagnosing CE compared to CD138 immunohistochemistry examination varied from 35.2% to 98.4%, specificity from 54.01% to 91.7%, PPV from 34.8% to 88.8%, and NPV from 57.5% to 97.8%. 13 , 14 , 15 In our study, the sensitivity of HSC in diagnosing CE was 47.5%, specificity was 64.38%, PPV was 50%, and NPV was 62.05% ( Table 4 ). Table 4 Performance of liquid hysteroscopy in the diagnosis of CE Table 4 Study Methodology and study year Population No Hysteroscopic diagnostic criteria Se (%) Sp (%) PPV (%) NPV (%) Yang et al 15 , a Retrospective, controlled, Part 1: May 2010 to April 2012. Part 2: Jan 2009 to Jan 2010. Part 1: 202 patients. Part 2: 60 patients with repeated implantation failures 2 —Micro polyps —Stromal edema —Hyperemia 35.2 67.5 45.6 57.5 Bouet et al 13 , a Prospective, Nov 2012 to March 2015 94 patients with recurrent pregnancy loss (43) and repeated implantation failures (51). 10 —Pedunculated and vascularized micro polyps <1 mm —Focal or diffuse hyperemia —Strawberry appearance 40 79.7 34.8 83.1 Song et al 16 , a Retrospective, Sep 2016 to Sep 2017 1189 patients in perimenopause. 2 —Diffuse or sporadic micro polyps —Hyperemia —Stromal edema 59.3 69.7 42.1 82.8 Liu et al 11 , a Prospective, Feb 2011 to June 2018 320 patients with unexplained infertility or recurrent pregnancy loss. 2 —Hemorrhagic spots —Focal or diffuse hyperemia —Micro polyps —Endometrial polyps —Dilated endometrial vessels 62.8 91.7 88.8 70.1 Mitter et al 14 , a Prospective, cohort Jan 2014 to Dec 2018 67 patients with implantation failures and recurrent pregnancy loss. 1 —Strawberry appearance —Stromal edema —Irregular endometrium —Hyperemia with predominant leucoplakia 42.8 76 - - Our study, 2022 a Prospective 70 infertile patients. 2 —Micro polyps <1 mm —Stromal edema —Focal or diffuse hyperemia —Hemorrhagic spots —Strawberry appearance 47.5 64.38 50 62.05 No , number of observers; NPV , negative predictive value; PPV , positive predictive value; Se , sensitivity; Sp , specificity. a Study uses IHC examination with the CD-138 antibody. Halouani. Performance of hysteroscopy in diagnosing chronic endometritis and the role of intra- and inter-observer variability. AJOG Glob Rep 2025. Performance of liquid hysteroscopy in the diagnosis of CE No , number of observers; NPV , negative predictive value; PPV , positive predictive value; Se , sensitivity; Sp , specificity. Study uses IHC examination with the CD-138 antibody. In our study, the diagnostic performance of HSC in detecting CE was lower than that reported in some previous studies. Several factors may explain this discrepancy. First, HSC is a dynamic examination, and the diagnosis of CE can be challenging, as endometrial signs are often subtle, focal, or difficult to interpret. In many studies, however, only static hysteroscopic images are presented, which may not accurately capture the complexity of real-time diagnosis. Second, unlike studies that relied on standard histopathology as the reference, we used endometrial biopsy combined with immunohistochemistry (CD138 staining), which is currently regarded as the most reliable method for diagnosing CE. Since CE is fundamentally a histopathological condition, immunohistochemistry is considered the gold standard for diagnosis. The relatively modest diagnostic performance of HSC observed in our study and reported in the literature suggests that, in patients with normal findings on transvaginal ultrasound and hysterosonography, it may be more accurate—and certainly more cost-effective—to proceed directly with endometrial biopsy and IHC rather than relying on HSC. Although our study focused on diagnostic accuracy, future research should assess how improved diagnosis and management of CE impact clinical outcomes, particularly fertility and pregnancy rates, in infertile women. To reduce inter-observer variability, Cicinelli et al 8 published unified criteria for the diagnosis of CE in 2019. They demonstrated strong agreement among the 62 observers in group A who used Cicinelli’s standardized criteria (ICC=0.78) vs poor agreement among the observers in group B who did not use these criteria (ICC=0.40). In this series, observers interpreted static digital hysteroscopic images. 8 A retrospective study published in 2018 showed strong agreement between two observers ( κ =0.73) and almost perfect intra-observer agreement ( κ =0.86). 16 In our study, inter-observer agreement was strong ( κ =0.62), and intra-observer agreement was moderate ( κ =0.58), despite the use of Cicinelli’s standardized criteria ( Table 4 ).

Infertility is described as the failure to achieve a pregnancy after at least a year of consistent, unprotected intercourse. Women older than 35 are advised to seek medical evaluation and treatment if they have not conceived after 6 months of trying. 1 Chronic endometritis (CE) is defined as a chronic inflammatory condition of the endometrium, histologically characterized by the presence of a plasma cell infiltrate within the endometrial mucosa. 2 , 3 It is often implicated in female infertility. 4 Clinical diagnosis of this condition is challenging due to the lack of specific symptoms. Histological examination with immunohistochemical (IHC) analysis of the plasma cell marker CD138 is the gold standard for diagnosing CE. 5 , 6 Hysteroscopic criteria for diagnosing CE can have a sensitivity of up to 98.4%. 7 These criteria were proposed by Cicinelli et al 8 in 2019 and are subject to intra- and inter-observer variability. We evaluated the performance of hysteroscopy (HSC) in diagnosing CE and assessed intra- and inter-observer variability in the hysteroscopic diagnosis of CE using the standardized criteria of Cicinelli et al. 8