Conventional microscopy versus digital image analysis for histopathologic evaluation of immune cells in the endometrium

AI-algorithm training and validation for identification of endometrial CD138+ cells in infertility-associated conditions; polycystic ovary syndrome (PCOS) and recurrent implantation failure (RIF)

Background

Endometrial CD138+ plasma cells serve as a diagnostic biomarker for endometrial inflammation, and their elevated occurrence correlates positively with adverse pregnancy outcomes. Infertility-related conditions like polycystic ovary syndrome (PCOS) and recurrent implantation failure (RIF) are closely associated with systemic and local chronic inflammatory status, wherein endometrial CD138+ plasma cell accumulation could also contribute to endometrial pathology. Current methods for quantifying CD138+ cells typically involve laborious and time-consuming microscopic assessments of only a few random areas from a slide. These methods have limitations in accurately representing the entire slide and are susceptible to significant biases arising from intra- and interobserver variations. Implementing artificial intelligence (AI) for CD138+ cell identification could enhance the accuracy, reproducibility, and reliability of analysis.

Methods

Here, an AI algorithm was developed to identify CD138+ plasma cells within endometrial tissue. The AI model comprised two layers of convolutional neural networks (CNNs). CNN1 was trained to segment epithelium and stroma across 28,363 mm2 (2.56 mm2 of epithelium and 24.87 mm2 of stroma), while CNN2 was trained to distinguish stromal cells based on CD138 staining, encompassing 7345 cells in the object layers (6942 CD138- cells and 403 CD138+ cells). The training and performance of the AI model were validated by three experienced pathologists. We collected 193 endometrial tissues from healthy controls (n = 73), women with PCOS (n = 91), and RIF patients (n = 29) and compared the CD138+ cell percentages based on cycle phases, ovulation status, and endometrial receptivity utilizing the AI model.

Results

The AI algorithm consistently and reliably distinguished CD138- and CD138+ cells, with total error rates of 6.32% and 3.23%, respectively. During the training validation, there was a complete agreement between the decisions made by the pathologists and the AI algorithm, while the performance validation demonstrated excellent accuracy between the AI and human evaluation methods (intraclass correlation; 0.76, 95% confidence intervals; 0.36-0.93, p = 0.002) and a positive correlation (Spearman's rank correlation coefficient: 0.79, p < 0.01). In the AI analysis, the AI model revealed higher CD138+ cell percentages in the proliferative phase (PE) endometrium compared to the secretory phase or anovulatory PCOS endometrium, irrespective of PCOS diagnosis. Interestingly, CD138+ percentages differed according to PCOS phenotype in the PE (p = 0.03). On the other hand, the receptivity status had no impact on the cell percentages in RIF samples.

Conclusion

Our findings emphasize the potential and accuracy of the AI algorithm in detecting endometrial CD138+ plasma cells, offering distinct advantages over manual inspection, such as rapid analysis of whole slide images, reduction of intra- and interobserver variations, sparing the valuable time of trained specialists, and consistent productivity. This supports the application of AI technology to help clinical decision-making, for example, in understanding endometrial cycle phase-related dynamics, as well as different reproductive disorders.

Endometrial Staining of CD56 (Uterine Natural Killer), BCL-6, and CD138 (Plasma Cells) Improve Diagnosis and Clinical Pregnancy Outcomes in Unexplained Infertility and Recurrent IVF Failures: Standardization of Diagnosis with Digital Pathology

In women with unexplained infertility (UI) and recurrent in vitro fertilization (IVF) failures, the etiology is often unclear. Endometrial immune perturbations and the use of immune markers associated with these dysregulations are of great interest in the diagnosis and treatment of UI. However, reliable biomarkers and standardized quantification methods are lacking. Here, to address endometrial immune dysregulation in UI patients with recurrent IVF failures, we performed endometrial tissue sampling and immunostaining of CD56 (uNK), CD138, and BCL-6. Of these cases, 57.9% had positive CD56 in the endometrial stroma, while 46.1% had positive BCL-6 in the glandular epithelium, and 14.5% of the cases were found to be positive for CD138. Combined staining rates were 60.5%, 68.4%, and 71.05% for (CD56 or BCL-6), (CD56 or CD138), and (CD56, BCL-6, or CD138), respectively. There was a significant correlation between CD56 and BCL-6 positivity, while CD138 positivity was an independent parameter. After the recommended targeted therapy, pregnancy rates were found to increase from 58.5% to 61.6% and 73.8% in CD56-positive, (CD56- or BCL-6-positive), and (CD56-, BCL-6-, or CD138-positive) cases, respectively. Notably, a retrospective evaluation of digital pathology and light microscopy results showed a significant correlation. This study suggests that the examination of CD56, BCL-6, and CD138 in the same endometrial sample may be an effective method in determining the etiology of UI and reaching an early diagnosis and treatment options. Moreover, digital pathology can be used in the evaluation of CD56 and BCL-6 to provide objective, rapid, and reliable results.

Endometrial HLA-F expression is influenced by genotypes and correlates differently with immune cell infiltration in IVF and recurrent implantation failure patients

STUDY QUESTION

Is human leukocyte antigen (HLA)-F protein expressed in mid-secretory endometrium, and are its expression levels influenced by HLA-F gene polymorphisms and correlated with the abundance of uterine natural killer (uNK) cells and anti-inflammatory M2 macrophages?

SUMMARY ANSWER

HLA-F protein is expressed in mid-secretory endometrium, and levels are correlated with immune cell infiltration, plasma progesterone concentrations and HLA-F single-nucleotide polymorphisms (SNPs), however, women experiencing recurrent implantation failure (RIF) show differences when compared to women attending their first IVF treatment.

WHAT IS KNOWN ALREADY

The immunomodulatory HLA class Ib molecules HLA-G and HLA-F are expressed on the extravillous trophoblast cells and interact with receptors on maternal immune cells. Little is known regarding HLA-F expression in endometrial stroma and HLA-F function; furthermore, HLA-F and HLA-G SNP genotypes and haplotypes have been correlated with differences in time-to-pregnancy.

STUDY DESIGN, SIZE, DURATION

Primary endometrial stromal cell (ESC) cultures (n = 5) were established from endometrial biopsies from women attending IVF treatment at a fertility clinic. Basic HLA-F and HLA-G protein expression by the ESCs were investigated. A prospective controlled cohort study was performed including 85 women with a history of RIF and 36 control women beginning their first fertility treatment and with no history of RIF. In some analyses, the RIF group was divided into unknown cause, male infertility, female infertility, and both female and male infertility. Endometrial biopsies and blood samples were obtained the day equivalent to embryo transfer in a hormone-substituted cycle.

PARTICIPANTS/MATERIALS, SETTING, METHODS

HLA protein expression by ESCs was characterized using flow cytometry and western blot. In the cohort study, the specific immune markers HLA-F and HLA-G, CD56 and CD16 (NK cells), CD163 (M2 macrophages), FOXP3 (regulatory T cells) and CD138 (plasma cells) were analysed by immunohistochemistry and a digital image analysis system in endometrial biopsies. Endometrial receptivity was assessed by an endometrial receptivity array test (the ERA® test). Endometrial biopsies were examined according to modified Noyes' criteria. SNPs at the HLA-F gene and HLA-G haplotypes were determined.

MAIN RESULTS AND THE ROLE OF CHANCE

HLA-F protein is expressed in the endometrium at the time of implantation. Furthermore, the HLA-F protein levels were different according to the womeńs HLA-F SNP genotypes and diplotypes, which have previously been correlated with differences in time-to-pregnancy. Endometrial HLA-F was positively correlated with anti-inflammatory CD163+ M2 macrophage infiltration and CD56+ uNK cell abundance for the entire cohort. However, this was not the case for CD56+ in the female infertility RIF subgroup. HLA-F levels in the endometrial stroma were negatively correlated with plasma progesterone concentrations in the RIF subgroup with known female infertility. Conversely, HLA-F and progesterone were positively correlated in the RIF subgroup with infertility of the male partner and no infertility diagnosis of the woman indicating interconnections between progesterone, HLA-F and immune cell infiltration. Glandular sHLA-G expression was also positively correlated with uNK cell abundance in the RIF subgroup with no female infertility but negatively correlated in the RIF subgroup with a female infertility diagnosis.

LARGE SCALE DATA

Immunohistochemistry analyses of endometrial biopsies and DNA sequencing of HLA genes. Data will be shared upon reasonable request to the corresponding author.

LIMITATIONS, REASONS FOR CAUTION

The control group of women attending their first IVF treatment had an anticipated good prognosis but was not proven fertile. A significant age difference between the RIF group and the IVF group reflects the longer treatment period for women with a history of RIF. The standardization of hormonal endometrial preparation, which allowed consistent timing of endometrial and blood sampling, might be a strength because a more uniform hormonal background may more clearly show an influence on the immune marker profile and HLA class Ib levels in the endometrium by other factors, for example genetic polymorphisms. However, the immune marker profile might be different during a normal cycle.

WIDER IMPLICATIONS OF THE FINDINGS

The findings further highlight the importance of HLA-F and HLA-G at the implantation site and in early pregnancy for pregnancy success. Diagnostic measures and modulation of the complex interactions between HLA class Ib molecules, maternal immune cells and hormonal factors may have potential to improve fertility treatment.

STUDY FUNDING/COMPETING INTEREST(S)

This work was supported by the Region Zealand Health Sciences Research Foundation and the Zealand University Hospital through the ReproHealth Research Consortium ZUH. The authors declared there are no conflicts of interest.