Model application to quantitatively evaluate placental features from ultrasound images with gestational diabetes.

PURPOSE: The goal of this study was to introduce PFCnet (placental features classification network), an multimodel model for evaluating and classifying placental features in gestational diabetes mellitus (GDM) and normal late pregnancy. Deep learning algorithms could be utilized to fully automate the examination of alterations in the placenta caused by hyperglycemia. METHODS: A total of 718 placental ultrasound images, including 139 cases of GDM, were collected, including gray-scale images (GSIs) and microflow images (MFIs). Ultrasonic assessment parameters and perinatal features were recorded. We divided gestational age into two categories for analysis (37 weeks and 37 weeks) based on the cut-off value level of placental maturity. The PFCnet model was introduced for identifying placental characteristics from normal and GDM pregnancies after extensive training and optimization. The model was scored using metrics such as sensitivity, specificity, accuracy, and the area under the curve (AUC). RESULTS: In view of multimodal fusion (GSIs and MFIs) and deep network optimization training, the overall diagnostic performance of the PFCnet model depending on the region of interest (ROI) was excellent (AUC: 93%), with a sensitivity of 89%, a specificity of 92%, and an accuracy of 92% in the independent test set. The fusion features of GSIs and MFIs in the placenta showed a higher discriminative power than single-mode features (accuracy: Fusion 92% vs. GSIs 84% vs. MFIs 82%). The independent test set at 37 weeks exhibited a better specificity (75% vs. 69%) but a lower sensitivity(95% vs. 100%). CONCLUSIONS: With its dual channel identification of placental parenchymal and vascular lesions in obstetric complications, the PFCnet classification model has the potential to be a useful tool for detecting placental tissue abnormalities caused by hyperglycemia.

Multimodal fusion model for classifying placenta ultrasound imaging in pregnancies with hypertension disorders.

BACKGROUND: A multimodal fusion model was proposed to assist the traditional visual diagnosis in evaluating the placental features of hypertension disorders of pregnancy (HDP). OBJECTIVE: The aim of this study was to analyse and compare the placental features between normal and HDP pregnancies and propose a multimodal fusion deep learning model for differentiating and characterizing the placental features from HDP to normal pregnancy. METHODS: This observational prospective study included 654 pregnant women, including 75 with HDPs. Grayscale ultrasound images (GSIs) and Microflow images (MFIs) of the placentas were collected from all patients during routine obstetric examinations. On the basis of intelligent extraction and features fusion, after quantities of training and optimization, the classification model named GMNet (the intelligent network based on GSIs and MFIs) was introduced for differentiating the placental features of normal and HDP pregnancies. The distributions of placental features extracted by the deep convolutional neural networks (DCNNs) were visualized by Uniform Manifold Approximation and Projection for Dimension Reduction (UMAP). Metrics including sensitivity, specificity, accuracy, and the area under the curve (AUC) were used to score the model. Finally, placental tissue samples were randomly selected for microscopic analyses to prove the interpretability and effectiveness of the GMNet model. RESULTS: Compared with the Normal group in ultrasonic images, the light spots were rougher and the parts with focal cystic or hypoechogenic lesions were increased in the HDP groups. The overall diagnostic performance of the GMNet model depending on the region of interest (ROI) was excellent (AUC: 97%), with a sensitivity of 90.0%, a specificity of 93.5%, and an accuracy of 93.1%. The fusion features of GSIs and MFIs in the placenta showed a higher discriminative power than single-mode features (fusion features vs GSI features vs MFI features, 97.0% vs 91.2% vs 94.8%). Furthermore, according to the microscopic analysis, unevenly distributed villi, increased syncyte nodules and aggregated intervillous cellulose deposition were particularly frequent in the HDP cases. CONCLUSIONS: The GMNet model could sensitively identify abnormal changes in the placental microstructure in pregnancies with HDP.

GMRLNet: A Graph-Based Manifold Regularization Learning Framework for Placental Insufficiency Diagnosis on Incomplete Multimodal Ultrasound Data.

Multimodal analysis of placental ultrasound (US) and microflow imaging (MFI) could greatly aid in the early diagnosis and interventional treatment of placental insufficiency (PI), ensuring a normal pregnancy. Existing multimodal analysis methods have weaknesses in multimodal feature representation and modal knowledge definitions and fail on incomplete datasets with unpaired multimodal samples. To address these challenges and efficiently leverage the incomplete multimodal dataset for accurate PI diagnosis, we propose a novel graph-based manifold regularization learning (MRL) framework named GMRLNet. It takes US and MFI images as input and exploits their modality-shared and modality-specific information for optimal multimodal feature representation. Specifically, a graph convolutional-based shared and specific transfer network (GSSTN) is designed to explore intra-modal feature associations, thus decoupling each modal input into interpretable shared and specific spaces. For unimodal knowledge definitions, graph-based manifold knowledge is introduced to describe the sample-level feature representation, local inter-sample relations, and global data distribution of each modality. Then, an MRL paradigm is designed for inter-modal manifold knowledge transfer to obtain effective cross-modal feature representations. Furthermore, MRL transfers the knowledge between both paired and unpaired data for robust learning on incomplete datasets. Experiments were conducted on two clinical datasets to validate the PI classification performance and generalization of GMRLNet. State-of-the-art comparisons show the higher accuracy of GMRLNet on incomplete datasets. Our method achieves 0.913 AUC and 0.904 balanced accuracy (bACC) for paired US and MFI images, as well as 0.906 AUC and 0.888 bACC for unimodal US images, illustrating its application potential in PI CAD systems.

Correlation between cerebroplacental doppler ratio and neonatal respiratory disorders: A reference marker of fetal lung maturation.

AIM: The aim of this study was to investigate the role of cerebroplacental ratio (CPR) in the final prenatal care for neonatal respiratory diseases and to analyze the risk of relevant factors associated with neonatal respiratory disorders. METHODS: A prospective cohort study of 795 singleton pregnancies was conducted. The pulsatility indices (PI) of the umbilical artery (UA) and the middle cerebral artery (MCA) were measured, and the MCA to UA ratio (CPR) was determined. The severity of the case is determined by whether or not the newborn has respiratory problems. Compare the CPR correlation between the two groups and examine the illness prediction factors through a binary logistic regression method. RESULTS: Of the 801 participants, 114 had neonatal respiratory disorders. The mean values of CPR between neonatal respiratory diseases group and control group were 1.78±0.6, 1.97±0.9, respectively (P <  0.001). Maternal age, abortion history, cesarean section history, placental thickness, placental maturity, and amniotic fluid index (AFI) were determined to have no significant link between the two groups after comparison analysis (P >  0.05). It could be found that compared with the control group, CPR MoM indicators of neonatal respiratory distress syndrome, neonatal pneumonia and wet lung disease all show significant decreases. In binary logistic regression analysis, among the variables included in the model, CPR (OR:2.90, P = 0.015), fetal heart monitoring (OR:5.26, P <  0.001), delivery mode (OR:2.86, P <  0.001) and gestational age of delivery (OR:0.92, P <  0.001) were statistically significant in both groups. CONCLUSION: The findings of this study showed that infant respiratory problems were substantially related to CPR value. The correlation indicates that CPR was a powerful reference marker for respiratory disorders.

Mitochondrial Autophagy and NLRP3 Inflammasome in Pulmonary Tissues from Severe Combined Immunodeficient Mice after Cardiac Arrest and Cardiopulmonary Resuscitation.

BACKGROUND: The incidence of cancer, diabetes, and autoimmune diseases has been increasing. Furthermore, there are more and more patients with solid organ transplants. The survival rate of these immunocompromised individuals is extremely low when they are severely hit-on. In this study, we established cardiac arrest cardiopulmonary resuscitation (CPR) model in severe combined immunodeficient (SCID) mice, analyzed the expression and activation of mitochondrial autophagy and NLRP3 inflammasome/caspase-1, and explored mitochondrial repair and inflammatory injury in immunodeficiency individual during systemic ischemia-reperfusion injury. METHODS: A potassium chloride-induced cardiac arrest model was established in C57BL/6 and nonobese diabetic/SCID (NOD/SCID) mice. One hundred male C57BL/6 mice and 100 male NOD/SCID mice were randomly divided into five groups (control, 2 h post-CPR, 12 h post-CPR, 24 h post-CPR, and 48 h post-CPR). A temporal dynamic view of alveolar epithelial cells, macrophages, and neutrophils from bronchoalveolar lavage fluid (BALF) was obtained using Giemsa staining. Spatial characterization of phenotypic analysis of macrophages in the lung interstitial tissue was analyzed by flow cytometry. The morphological changes of mitochondria 48 h after CPR were studied by transmission electron microscopy and quantified according to the Flameng grading system. Western blotting analysis was used to detect the expression and activation of the markers of mitochondrial autophagy, NLRP3 inflammasome, and caspase-1. RESULTS: (1) In NOD/SCID mice, macrophages were disintegrated in BALF, and many alveolar epithelial cells were shed at 48 h after resuscitation. Compared with C57BL/6 mice, the ratio of macrophages/total cells peaked at 12 h and was significantly higher in NOD/SCID mice (31.17 ± 4.13 vs. 49.69 ± 2.43, t = 14.46, P = 0.001). After 24 h, the results showed a downward trend. Furthermore, a large number of macrophages were disintegrated in the BALF. (2) Mitochondrial autophagy was present in both C57BL/6 and NOD/SCID mice after CPR, but it began late in the NOD/SCID mice. Compared with C57BL/6 mice, phos-ULK1 (Ser327) expression was significantly lower at 2 h and 12 h after CPR (2 h after CPR: 1.88 ± 0.36 vs. 1.12 ± 0.11, t = -1.36, P < 0.01 and 12 h after CPR: 1.52 ± 0.16 vs. 1.05 ± 0.12, t = -0.33, P < 0.01), whereas phos-ULK1 (Ser757) expression was significantly higher at 2 h and 12 h after CPR in NOD/SCID mice (2 h after CPR: 1.28 ± 0.12 vs. 1.69 ± 0.14, t = 1.7, P < 0.01 and 12 h after CPR: 1.33 ± 0.10 vs. 1.94 ± 0.13, t = 2.75, P < 0.01). (3) Furthermore, NLRP3 inflammasome/caspase-1 activation in the pulmonary tissues occurred early and for only a short time in C57BL/6 mice, but this phenomenon was sustained in NOD/SCID mice. The expression of the NLRP3 inflammasome increased modestly in the C57 mice, but the increase was higher in the NOD/SCID mice than in the C57BL/6 mice, especially at 12, 24, 48 h after CPR (48 h after CPR: 1.46 ± 0.13 vs. 2.97 ± 0.19, t = 5.34, P = 0.001). The expression of caspase-1-20 generally followed the same pattern as the NLRP3 inflammasome. CONCLUSIONS: There is a regulatory relationship between the NLRP3 inflammasome and mitochondrial autophagy after CPR in the healthy mice. This regulatory relationship was disturbed in the NOD/SCID mice because the signals for mitochondrial autophagy occurred late, and NLRP3 inflammasome- and caspase-1-dependent cell injury was sustained.