Automated image analysis of placental villi and syncytial knots in histological sections

Characterising delayed villous maturation: A narrative literature review

The normal development of the placenta is vital for fetal growth and a healthy pregnancy outcome. Delayed villous maturation (DVM) is a placental lesion that has been implicated in stillbirth. In DVM, villi do not maturate adequately for their gestational age. DVM is characterised by larger and fewer terminal placental villi, low numbers of syncytial knots, and thicker and fewer vasculosyncytial membranes. DVM is most commonly reported in conjunction with maternal diabetes; however, the occurrence of idiopathic DVM suggests that there may be multiple mechanistic pathways that contribute to DVM. DVM can only be diagnosed through histopathological examination after birth, and there is significant interobserver variability in diagnosis. Establishing objective criteria to distinguish between DVM and healthy placentas is key to increasing the understanding of DVM. Vasculosyncytial membrane count, numbers of syncytial knots and CD15, among others, have been presented as potential diagnostic criteria in the literature. This review aims to compile information on DVM, including the pathophysiology, conditions that have reported associations with DVM and potential markers that could be used as criteria to differentiate between DVM and healthy placentas.

Inhibition of NFAT promotes loss of tissue resident uterine natural killer cells and attendant pregnancy complications in humans

Uterine natural killer cells (uNKs) are a tissue resident lymphocyte population that are critical for pregnancy success. Although mouse models have demonstrated that NK deficiency results in abnormal placentation and poor pregnancy outcomes, the generalizability of this knowledge to humans remains unclear. Here we identify uterus transplant (UTx) recipients as a human population with reduced endometrial NK cells and altered pregnancy phenotypes. We further show that the NK reduction in UTx is due to impaired transcriptional programming of NK tissue residency due to blockade of the transcription factor nuclear factor of activated T cells (NFAT). NFAT-dependent genes played a role in multiple molecular circuits governing tissue residency in uNKs, including early residency programs involving AP-1 transcription factors as well as TGFβ-mediated upregulation of surface integrins. Collectively, our data identify a previously undescribed role for NFAT in uterine NK tissue residency and provide novel mechanistic insights into the biologic basis of pregnancy complications due to alteration of tissue resident NK subsets in humans.

One Sentence Summary

Role of NFAT in uterine NK cell tissue residency.

Tissue Contamination Challenges the Credibility of Machine Learning Models in Real World Digital Pathology

Machine learning (ML) models are poised to transform surgical pathology practice. The most successful use attention mechanisms to examine whole slides, identify which areas of tissue are diagnostic, and use them to guide diagnosis. Tissue contaminants, such as floaters, represent unexpected tissue. Although human pathologists are extensively trained to consider and detect tissue contaminants, we examined their impact on ML models. We trained 4 whole-slide models. Three operate in placenta for the following functions: (1) detection of decidual arteriopathy, (2) estimation of gestational age, and (3) classification of macroscopic placental lesions. We also developed a model to detect prostate cancer in needle biopsies. We designed experiments wherein patches of contaminant tissue are randomly sampled from known slides and digitally added to patient slides and measured model performance. We measured the proportion of attention given to contaminants and examined the impact of contaminants in the t-distributed stochastic neighbor embedding feature space. Every model showed performance degradation in response to one or more tissue contaminants. Decidual arteriopathy detection--balanced accuracy decreased from 0.74 to 0.69 ± 0.01 with addition of 1 patch of prostate tissue for every 100 patches of placenta (1% contaminant). Bladder, added at 10% contaminant, raised the mean absolute error in estimating gestational age from 1.626 weeks to 2.371 ± 0.003 weeks. Blood, incorporated into placental sections, induced false-negative diagnoses of intervillous thrombi. Addition of bladder to prostate cancer needle biopsies induced false positives, a selection of high-attention patches, representing 0.033 mm2, and resulted in a 97% false-positive rate when added to needle biopsies. Contaminant patches received attention at or above the rate of the average patch of patient tissue. Tissue contaminants induce errors in modern ML models. The high level of attention given to contaminants indicates a failure to encode biological phenomena. Practitioners should move to quantify and ameliorate this problem.

Tissue contamination challenges the credibility of machine learning models in real world digital pathology

Machine learning (ML) models are poised to transform surgical pathology practice. The most successful use attention mechanisms to examine whole slides, identify which areas of tissue are diagnostic, and use them to guide diagnosis. Tissue contaminants, such as floaters, represent unexpected tissue. While human pathologists are extensively trained to consider and detect tissue contaminants, we examined their impact on ML models. We trained 4 whole slide models. Three operate in placenta for 1) detection of decidual arteriopathy (DA), 2) estimation of gestational age (GA), and 3) classification of macroscopic placental lesions. We also developed a model to detect prostate cancer in needle biopsies. We designed experiments wherein patches of contaminant tissue are randomly sampled from known slides and digitally added to patient slides and measured model performance. We measured the proportion of attention given to contaminants and examined the impact of contaminants in T-distributed Stochastic Neighbor Embedding (tSNE) feature space. Every model showed performance degradation in response to one or more tissue contaminants. DA detection balanced accuracy decreased from 0.74 to 0.69 +/- 0.01 with addition of 1 patch of prostate tissue for every 100 patches of placenta (1% contaminant). Bladder, added at 10% contaminant raised the mean absolute error in estimating gestation age from 1.626 weeks to 2.371 +/ 0.003 weeks. Blood, incorporated into placental sections, induced false negative diagnoses of intervillous thrombi. Addition of bladder to prostate cancer needle biopsies induced false positives, a selection of high-attention patches, representing 0.033mm2, resulted in a 97% false positive rate when added to needle biopsies. Contaminant patches received attention at or above the rate of the average patch of patient tissue. Tissue contaminants induce errors in modern ML models. The high level of attention given to contaminants indicates a failure to encode biological phenomena. Practitioners should move to quantify and ameliorate this problem.

Application of Digital Imaging and Artificial Intelligence to Pathology of the Placenta

Digital imaging, including the use of artificial intelligence, has been increasingly applied to investigate the placenta and its related pathology. However, there has been no comprehensive review of this body of work to date. The aim of this study was to therefore review the literature regarding digital pathology of the placenta. A systematic literature search was conducted in several electronic databases. Studies involving the application of digital imaging and artificial intelligence techniques to human placental samples were retrieved and analyzed. Relevant articles were categorized by digital image technique and their relevance to studying normal and diseased placenta. Of 2008 retrieved articles, 279 were included. Digital imaging research related to the placenta was often coupled with immunohistochemistry, confocal microscopy, 3D reconstruction, and/or deep learning algorithms. By significantly increasing pathologists' ability to recognize potentially prognostic relevant features and by lessening inter-observer variability, published data overall indicate that the application of digital pathology to placental and perinatal diseases, along with clinical and radiology correlation, has great potential to improve fetal and maternal health care including the selection of targeted therapy in high-risk pregnancy.

Automated Quantitative Image Evaluation of Antigen Retrieval Methods for 17 Antibodies in Placentation and Implantation Diagnostic and Research

Immunostaining in clinical routine and research highly depends on standardized staining methods and quantitative image analyses. We qualitatively and quantitatively compared antigen retrieval methods (no pretreatment, pretreatment with pepsin, and heat-induced pretreatment with pH 6 or pH 9) for 17 antibodies relevant for placenta and implantation diagnostics and research. Using our newly established, comprehensive automated quantitative image analysis approach, fluorescent signal intensities were evaluated. Automated quantitative image analysis found that 9 out of 17 antibodies needed antigen retrieval to show positive staining. Heat induction proved to be the most efficient form of antigen retrieval. Eight markers stained positive after pepsin digestion, with β-hCG and vWF showing enhanced staining intensities. To avoid the misinterpretation of quantitative image data, the qualitative aspect should always be considered. Results from native placental tissue were compared with sections of a placental invasion model based on thermo-sensitive scaffolds. Immunostaining on placentas in vitro leads to new insights into fetal development and maternal pathophysiological pathways, as pregnant women are justifiably excluded from clinical studies. Thus, there is a clear need for the assessment of reliable immunofluorescent staining and pretreatment methods. Our evaluation offers a powerful tool for antibody and pretreatment selection in placental research providing objective and precise results.

Association of endothelial nitric oxide synthase gene variants with preeclampsia

Background

Preeclampsia (PE) is a complex pregnancy hypertensive disorder with multifaceted etiology. The endothelial nitric oxide synthase (eNOS) gene and nitric oxide (NO) levels has been reported to be associated with PE predisposition in various populations. Therefore, present study was designed to investigate the role of NO levels and eNOS gene variants in preeclamptic women in Pakistan.

Methods

A total of 600 women were evaluated, 188 of PE with mild features, 112 of PE with severe features and 300 normotensive pregnant women. NO levels were detected by Greiss reaction method and genotyping following sequencing was conducted for eNOS gene variants. Further insilico studies were performed to get insights into the structural and functional impact of identifies mutation on eNOS protein as well as on protein regulation.

Results

Reduced concentrations of NO were reported in all PE groups (p < 0.05) as compared to controls. The frequency of c.894 T (p.298Asp) and g.-786C alleles were significantly associated with PE. In addition, novel homozygous variant g.2051G > A was also significantly associated with PE when compared to normotensive women. Dynamic simulation studies revealed that Glu298Asp mutation destabilize the protein molecule and decrease the overall stability of eNOS protein. Molecular docking analysis of mutant promoter with transcription factors STAT3 and STAT6 proposed changes in protein regulation upon these reported mutations in upstream region of the gene.

Conclusion

Considering the results of current study, the functional alterations induced by these variants may influence the bioavailability of NO and represents a genetic risk factor for increased susceptibility to PE. However, large studies or meta-analysis are necessary to validate these findings.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12978-021-01213-9.

Preeclampsia (PE) is a complex pregnancy hypertensive disorder with multifaceted etiology characterized by increased hypertension and proteinuria after 20 weeks of gestation. The present study was directed to determine the role of eNOS in susceptibility to PE and the association of c.894G > T (p.(Glu298Asp), intron 4b/4a, g.-786 T > C and other possible variants of eNOS gene with preeclampsia in Pakistani population. Computational analysis of identified variants in the coding and non-coding region of the eNOS gene was also conducted to determine the change in gene regulation and further protein stability. A total of 600 women were evaluated, 188 with mild and 112 with PE with severe features PE with 300 normotensive pregnant women. NO levels and genotyping following sequencing was conducted for eNOS gene variants. Further insilico studies were performed to get insights into the structural and functional impact of identifies mutation on eNOS protein as well as on protein regulation. Data from the current study suggest that there might be other risk variants of the eNOS gene (g.2051G > A and g.1861G > A) and lower levels of serum NO that confers in an increased risk of PE. The detailed computational investigation further confirmed the deformities and changes in protein flexibility upon Glu298Asp. These structural alterations might be associated with preeclampsia. Variants in the promoter region of the eNOS gene further validate the change in gene regulation for the onset of disease. Identification of key structural and functional features in eNOS protein and gene regulatory region might be used for designing specific drugs for therapeutic purpose.

New Frontiers in Placenta Tissue Imaging

The placenta is a highly vascularized organ with unique structural and metabolic complexities. As the primary conduit of fetal support, the placenta mediates transport of oxygen, nutrients, and waste between maternal and fetal blood. Thus, normal placenta anatomy and physiology is absolutely required for maintenance of maternal and fetal health during pregnancy. Moreover, impaired placental health can negatively impact offspring growth trajectories as well as increase the risk of maternal cardiovascular disease later in life. Despite these crucial roles for the placenta, placental disorders, such as preeclampsia, intrauterine growth restriction (IUGR), and preterm birth, remain incompletely understood. Effective noninvasive imaging and image analysis are needed to advance the obstetrician's clinical reasoning toolkit and improve the utility of the placenta in interpreting maternal and fetal health trajectories. Current paradigms in placental imaging and image analysis aim to improve the traditional imaging techniques that may be time-consuming, costly, or invasive. In concert with conventional clinical approaches such as ultrasound (US), advanced imaging modalities can provide insightful information on the structure of placental tissues. Herein we discuss such imaging modalities, their specific applications in structural, vascular, and metabolic analysis of placental health, and emerging frontiers in image analysis research in both preclinical and clinical contexts.

AI-PLAX: AI-based placental assessment and examination using photos

Post-delivery analysis of the placenta is useful for evaluating health risks of both the mother and baby. In the U.S., however, only about 20% of placentas are assessed by pathology exams, and placental data is often missed in pregnancy research because of the additional time, cost, and expertise needed. A computer-based tool that can be used in any delivery setting at the time of birth to provide an immediate and comprehensive placental assessment would have the potential to not only to improve health care, but also to radically improve medical knowledge. In this paper, we tackle the problem of automatic placental assessment and examination using photos. More concretely, we first address morphological characterization, which includes the tasks of placental image segmentation, umbilical cord insertion point localization, and maternal/fetal side classification. We also tackle clinically meaningful feature analysis of placentas, which comprises detection of retained placenta (i.e., incomplete placenta), umbilical cord knot, meconium, abruption, chorioamnionitis, and hypercoiled cord, and categorization of umbilical cord insertion type. We curated a dataset consisting of approximately 1300 placenta images taken at Northwestern Memorial Hospital, with hand-labeled pixel-level segmentation map, cord insertion point and other information extracted from the associated pathology reports. We developed the AI-based Placental Assessment and Examination system (AI-PLAX), which is a novel two-stage photograph-based pipeline for fully automated analysis. In the first stage, we use three encoder-decoder convolutional neural networks with a shared encoder to address morphological characterization tasks by employing a transfer-learning training strategy. In the second stage, we employ distinct sub-models to solve different feature analysis tasks by using both the photograph and the output of the first stage. We evaluated the effectiveness of our pipeline by using the curated dataset as well as the pathology reports in the medical record. Through extensive experiments, we demonstrate our system is able to produce accurate morphological characterization and very promising performance on aforementioned feature analysis tasks, all of which may possess clinical impact and contribute to future pregnancy research. This work is the first for comprehensive, automated, computer-based placental analysis and will serve as a launchpad for potentially multiple future innovations.

The effect of mother's β-thalassemia minor on placental histology and neonatal outcomes

Objective: β-thalassemia is a prevalent disease in Iran. The adverse effects of anemia on placental and neonatal outcomes have previously been shown. Women require additional iron during pregnancy and the anemia of women with β-thalassemia may adversely affect the neonatal outcome and increase placental abnormalities. In this study, we compared the placental histology and neonatal outcomes among pregnant women with and without β-thalassemia.Material and methods: In this population-based cross-sectional study, 144 pregnant women with β-thalassemia minor (case group) were compared to 142 women without β-thalassemia (control group). Women with singleton pregnancies over 20 weeks of gestation without pregnancy complications, anemia, collagen vascular diseases, or other hemoglobinopathies, all referred to hospitals affiliated to Shiraz University of Medical Sciences from March 2014 to February 2016, were included and maternal and neonatal data were extracted from medical records. After child birth, the placenta was evaluated for macroscopic and microscopic changes.Results: The frequency of LBW was significantly higher in the case group than that in the control group (19.7 vs. 9.7%, respectively) (p = .019). Six in the case group had gross abnormalities of placenta, while none in the control group (p = .03) and the case group had a higher frequency of chorangiosis, calcification, syncytial knot, and umbilical cord position (p < .05).Conclusion: As β-thalassemia increases the chance of placental abnormalities, it is recommended to pay more attention to patients with β-thalassemia, especially during pregnancy, to prevent placental and neonatal adverse outcomes.

The effect of maternal obstructive sleep apnea on the placenta

STUDY OBJECTIVES

Obstructive sleep apnea (OSA) during pregnancy has been associated with adverse maternal outcomes. However, the effect of maternal OSA on fetal growth is less clear. The placenta is a critical organ for fetal growth and development and the principal determinant of birthweight. We aimed to investigate the effect of maternal OSA on placental growth and function.

METHODS

Placentas of women recruited to a prospective longitudinal study were consecutively obtained immediately after delivery. Each placenta was measured for length, width, and thickness. Total RNA was isolated for gene expression analysis of VEGF, VEGF receptor, PIGF, and leptin. Histological and morphometric evaluations of the placenta were performed.

RESULTS

A total of 53 placentas were investigated. Ten women (19%) had OSA, and the weight of their placentas was significantly higher compared with the placentas of the controls (526.1 ± 83.9 vs. 425.7 ± 95.5 g, p = 0.004). There was a significant positive correlation between placental weight and the log apnea-hypopnea index even after controlling for maternal body mass index (BMI; r = 0.31, p = 0.04). The birthweight/placental weight ratio was significantly lower in women with OSA compared with controls (p = 0.03). Placental weight and newborn triceps adiposity thickness correlated positively after controlling for maternal BMI (r = 0.29, p = 0.04). Leptin expression was 1.8-fold higher in placentas of women with OSA compared with controls (p = 0.02). No histological differences were found between the groups.

CONCLUSIONS

Maternal OSA is associated with increased placental weight that correlated with OSA severity and neonatal adiposity independently of maternal BMI. Placental leptin overexpression may mediate/underlie the above findings.Trial Registration: Clinical Trials NCT00931099.

Automated segmentation of villi in histopathology images of placenta

PURPOSE

Manual analysis of clinical placenta pathology samples under the microscope is a costly and time-consuming task. Computer-aided diagnosis might offer a means to obtain fast and reliable results and also substantially reduce inter- and intra-rater variability. Here, we present a fully automated segmentation method that is capable of distinguishing the complex histological features of the human placenta (i.e., the chorionic villous structures).

METHODS

The proposed pipeline consists of multiple steps to segment individual placental villi structures in hematoxylin and eosin (H&E) stained placental images. Artifacts and undesired objects in the histological field of view are detected and excluded from further analysis. One of the challenges in our new algorithm is the detection and segmentation of touching villi in our dataset. The proposed algorithm uses the top-hat transformation to detect candidate concavities in each structure, which might represent two distinct villous structures in close proximity. The detected concavities are classified by extracting multiple features from each candidate concavity. Our proposed pipeline is evaluated against manual segmentations, confirmed by an expert pathologist, on 12 scans from three healthy control patients and nine patients diagnosed with preeclampsia, containing nearly 5000 individual villi. The results of our method are compared to a previously published method for villi segmentation.

RESULTS

Our algorithm detected placental villous structures with an F1 score of 80.76% and sensitivity of 82.18%. These values are substantially better than the previously published method, whose F1 score and sensitivity are 65.30% and 55.12%, respectively.

CONCLUSION

Our method is capable of distinguishing the complex histological features of the human placenta (i.e., the chorionic villous structures), removing artifacts over a large histopathology sample of human placenta, and (importantly) account for touching adjacent villi structures. Compared to existing methods, our developed method yielded high accuracy in detecting villi in placental images.

Role of ACE I/D polymorphism in pathological assessment of preeclampsia in Pakistan

BACKGROUND

Preeclampsia (PE) is a pregnancy-related hypertensive disorder, which may stem from impair placentation. Renin-angiotensin system is one of the mediators of decidualization and trophoblastic proliferation. In the present study women with PE were studied in a comparison of normotensive controls to determine whether Angiotensin-converting enzyme (ACE) gene I/D polymorphism affect the placental villi and umbilical cord formation with the assessment of biochemical and clinical risk factors.

METHODS

Total 400 blood (PE/controls = 200), 400 urine (PE/controls = 200), 90 tissue samples of UC (PE = 50, controls = 40) and 90 placental tissue samples (PE = 50, controls = 40) were recruited. Histomorphological and Histomorphometric analysis were done for placental and umbilical cord tissues. Blood and serum parameters were analyzed, samples were genotyped for I/D polymorphism. Data were statistically analyzed by Independent sample t-test, Chi-square test and the odds ratio.

RESULTS

Histological study revealed significant increase (p < 0.001) in distance from Wharton jelly (in both artery and vein) and outer layer thickness of vein; significant reduction (p < 0.01 and p < 0.05) in the lumen area of artery and vein. Abnormal villi, more syncytial knots (SK) and a significant decrease in elongated and large villi in PE placentas. Analysis of ACE gene determined that genotypic frequencies were statistically significant (p < 0.02) among both groups and DD genotype was predominant in the PE group.

CONCLUSION

Present study reveals that ACE I/D polymorphism might affect the normal placental villi and umbilical cord formation in women with PE. In addition, histological studies and genetic evaluation can provide useful information in the determination of various reasons and mechanisms involved in the pathogenesis of PE in Pakistan.

Interventions for investigating and identifying the causes of stillbirth

BACKGROUND

Identification of the causes of stillbirth is critical to the primary prevention of stillbirth and to the provision of optimal care in subsequent pregnancies. A wide variety of investigations are available, but there is currently no consensus on the optimal approach. Given their cost and potential to add further emotional burden to parents, there is a need to systematically assess the effect of these interventions on outcomes for parents, including psychosocial outcomes, economic costs, and on rates of diagnosis of the causes of stillbirth.

OBJECTIVES

To assess the effect of different tests, protocols or guidelines for investigating and identifying the causes of stillbirth on outcomes for parents, including psychosocial outcomes, economic costs, and rates of diagnosis of the causes of stillbirth.

SEARCH METHODS

We searched Cochrane Pregnancy and Childbirth's Trials Register (31 August 2017), ClinicalTrials.gov and the WHO International Clinical Trials Registry Platform (ICTRP) (15 May 2017).

SELECTION CRITERIA

We planned to include randomised controlled trials (RCTs), quasi-RCTs, and cluster-RCTs. We planned to include studies published as abstract only, provided there was sufficient information to allow us to assess study eligibility. We planned to exclude cross-over trials.Participants included parents (including mothers, fathers, and partners) who had experienced a stillbirth of 20 weeks' gestation or greater.This review focused on interventions for investigating and identifying the causes of stillbirth. Such interventions are likely to be diverse, but could include:* review of maternal and family history, and current pregnancy and birth history;* clinical history of present illness;* maternal investigations (such as ultrasound, amniocentesis, antibody screening, etc.);* examination of the stillborn baby (including full autopsy, partial autopsy or noninvasive components, such as magnetic resonance imaging (MRI), computerised tomography (CT) scanning, and radiography);* umbilical cord examination;* placental examination including histopathology (microscopic examination of placental tissue); and* verbal autopsy (interviews with care providers and support people to ascertain causes, without examination of the baby).We planned to include trials assessing any test, protocol or guideline (or combinations of tests/protocols/guidelines) for investigating the causes of stillbirth, compared with the absence of a test, protocol or guideline, or usual care (further details are presented in the Background, see Description of the intervention).We also planned to include trials comparing any test, protocol or guideline (or combinations of tests/protocols/guidelines) for investigating the causes of stillbirth with another, for example, the use of a limited investigation protocol compared with a comprehensive investigation protocol.

DATA COLLECTION AND ANALYSIS

Two review authors assessed trial eligibility independently.

MAIN RESULTS

We excluded five studies that were not RCTs. There were no eligible trials for inclusion in this review.

AUTHORS' CONCLUSIONS

There is currently a lack of RCT evidence regarding the effectiveness of interventions for investigating and identifying the causes of stillbirth. Seeking to determine the causes of stillbirth is an essential component of quality maternity care, but it remains unclear what impact these interventions have on the psychosocial outcomes of parents and families, the rates of diagnosis of the causes of stillbirth, and the care and management of subsequent pregnancies following stillbirth. Due to the absence of trials, this review is unable to inform clinical practice regarding the investigation of stillbirths, and the specific investigations that would determine the causes.Future RCTs addressing this research question would be beneficial, but the settings in which the trials take place, and their design, need to be given careful consideration. Trials need to be conducted with the utmost care and consideration for the needs, concerns, and values of parents and families. Assessment of longer-term psychosocial variables, economic costs to health services, and effects on subsequent pregnancy care and outcomes should also be considered in any future trials.