Association between gestational cardiovascular health in the first trimester and pregnancy outcomes in the China birth cohort

OBJECTIVES

To determine whether gestational cardiovascular health (CVH) during the first trimester is associated with a risk of adverse pregnancy outcomes.

STUDY DESIGN

A multicentre prospective cohort; part of the China birth cohort study.

METHODS

Pregnant women were recruited at 6-13+6 gestation weeks and followed to delivery to identify pregnancy outcomes. Gestational CVH in the first trimester was assessed using five CVH metrics: body mass index, smoking, blood pressure, glucose, and lipids. Multilevel modified Poisson regression models calculated the relative risks (RRs) and 95% confidence intervals (95% CIs) of gestational CVH for adverse pregnancy outcomes.

RESULTS

Among 56,852 pregnant women, the mean score for gestational CVH during the first trimester was 9.1. Adjusting for confounding factors, each 1-point decrease in the total gestational CVH score significantly increased the risk of hypertensive disorders of pregnancy (RR = 1.682, 95% CI: 1.624-1.743), gestational diabetes mellitus (RR = 1.405, 95% CI: 1.384-1.426), preterm birth (RR = 1.184, 95% CI: 1.174-1.195), large for gestational age (RR = 1.224, 95% CI: 1.199-1.250), caesarean delivery (RR = 1.073, 95% CI: 1.049-1.097), and low Apgar score (RR = 1.131, 95% CI: 1.003-1.277) significantly increased. Meanwhile, the risk of small for gestational age decreased (SGA; RR = 0.922, 95% CI: 0.898-0.946). Worsened CVH categories significantly increased the risk of adverse pregnancy outcomes, excluding SGA.

CONCLUSIONS

Poor gestational CVH in the first trimester significantly increases the risk of adverse pregnancy outcomes, emphasising the need for early improvement in gestational CVH.

Study of association between corneal shape parameters and axial length elongation during orthokeratology using image-pro plus software

BACKGROUND

The aim was to validate the correlation between corneal shape parameters and axial length growth (ALG) during orthokeratology using Image-Pro Plus (IPP) 6.0 software.

METHODS

This retrospective study used medical records of myopic children aged 8-13 years (n = 104) undergoing orthokeratology. Their corneal topography and axial length were measured at baseline and subsequent follow-ups after lens wear. Corneal shape parameters, including the treatment zone (TZ) area, TZ diameter, TZ fractal dimension, TZ radius ratio, eccentric distance, pupil area, and pupillary peripheral steepened zone(PSZ) area, were measured using IPP software. The impact of corneal shape parameters at 3 months post-orthokeratology visit on 1.5-year ALG was evaluated using multivariate linear regression analysis.

RESULTS

ALG exhibited significant associations with age, TZ area, TZ diameter, TZ fractal dimension, and eccentric distance on univariate linear regression analysis. Multivariate regression analysis identified age, TZ area, and eccentric distance as significantly correlated with ALG (all P < 0.01), with eccentric distance showing the strongest correlation (β = -0.370). The regressive equation was y = 1.870 - 0.235a + 0.276b - 0.370c, where y represents ALG, a represents age, b represents TZ area, and c represents eccentric distance; R2 = 0.27). No significant relationships were observed between the TZ radius ratio, pupillary PSZ area, and ALG.

CONCLUSIONS

IPP software proves effective in capturing precise corneal shape parameters after orthokeratology. Eccentric distance, rather than age or the TZ area, significantly influences ALG retardation.

Multimodal risk prediction with physiological signals, medical images and clinical notes

The broad adoption of electronic health record (EHR) systems brings us a tremendous amount of clinical data and thus provides opportunities to conduct data-based healthcare research to solve various clinical problems in the medical domain. Machine learning and deep learning methods are widely used in the medical informatics and healthcare domain due to their power to mine insights from raw data. When adapting deep learning models for EHR data, it is essential to consider its heterogeneous nature: EHR contains patient records from various sources including medical tests (e.g. blood test, microbiology test), medical imaging, diagnosis, medications, procedures, clinical notes, etc. Those modalities together provide a holistic view of patient health status and complement each other. Therefore, combining data from multiple modalities that are intrinsically different is challenging but intuitively promising in deep learning for EHR. To assess the expectations of multimodal data, we introduce a comprehensive fusion framework designed to integrate temporal variables, medical images, and clinical notes in EHR for enhanced performance in clinical risk prediction. Early, joint, and late fusion strategies are employed to combine data from various modalities effectively. We test the model with three predictive tasks: in-hospital mortality, long length of stay, and 30-day readmission. Experimental results show that multimodal models outperform uni-modal models in the tasks involved. Additionally, by training models with different input modality combinations, we calculate the Shapley value for each modality to quantify their contribution to multimodal performance. It is shown that temporal variables tend to be more helpful than CXR images and clinical notes in the three explored predictive tasks.

Global, regional, and national trends in osteoarthritis disability-adjusted life years (DALYs) from 1990 to 2019: a comprehensive analysis of the global burden of disease study

OBJECTIVES

This study aimed to investigate global, regional, and national trends in osteoarthritis disability-adjusted life years (DALYs) from 1990 to 2019, identify the burden of osteoarthritis in different age groups, and assess age, period, and cohort effects on osteoarthritis DALYs.

STUDY DESIGN

A comprehensive analysis of the Global Burden of Disease Study 2019 data, covering 204 countries and territories.

METHODS

We conducted a comprehensive analysis using data from the Global Burden of Disease Study 2019, encompassing 204 countries and territories. Age-standardized DALY rates were calculated, and the age-period-cohort model was employed to examine the age, period, and cohort effects on osteoarthritis DALYs. The annual percentage change (APC) and average annual percentage change (AAPC) were estimated to evaluate trends in DALYs.

RESULTS

Globally, osteoarthritis DALYs increased by 114.48 % between 1990 and 2019, with an age-standardized DALY rate growth of 3.3 %. The largest relative growth in DALYs occurred in Middle and Low-middle Socio-Demographic Index (SDI) regions. DALYs increased significantly in almost all age-specific groups, particularly among 45-74 years old age groups. Age, period, and cohort effects analysis revealed a general increase in osteoarthritis DALYs risk over time, with some variations by SDI quintiles and sex. The steepest increase in DALYs occurred in the 30-34 years age group, and the trend attenuated with increasing age. Males showed significantly slower DALYs growth than females in age groups with non-overlapping 95 % confidence intervals. Age effects were consistently higher in females, especially in high-SDI countries. Period and cohort effects generally demonstrated a climbing risk of osteoarthritis DALYs across different SDI quintiles, with more pronounced increases in lower-SDI regions.

CONCLUSIONS

Our findings highlight the substantial and increasing burden of osteoarthritis at global, regional, and national levels from 1990 to 2019, with significant variations by age, period, and cohort. These results underscore the importance of developing targeted public health strategies and interventions to address the growing impact of osteoarthritis, particularly in lower-SDI regions and among older populations.

Stable clinical risk prediction against distribution shift in electronic health records

The availability of large-scale electronic health record datasets has led to the development of artificial intelligence (AI) methods for clinical risk prediction that help improve patient care. However, existing studies have shown that AI models suffer from severe performance decay after several years of deployment, which might be caused by various temporal dataset shifts. When the shift occurs, we have access to large-scale pre-shift data and small-scale post-shift data that are not enough to train new models in the post-shift environment. In this study, we propose a new method to address the issue. We reweight patients from the pre-shift environment to mitigate the distribution shift between pre- and post-shift environments. Moreover, we adopt a Kullback-Leibler divergence loss to force the models to learn similar patient representations in pre- and post-shift environments. Our experimental results show that our model efficiently mitigates temporal shifts, improving prediction performance.

Multimodal Risk Prediction with Physiological Signals, Medical Images and Clinical Notes

The broad adoption of electronic health records (EHRs) provides great opportunities to conduct healthcare research and solve various clinical problems in medicine. With recent advances and success, methods based on machine learning and deep learning have become increasingly popular in medical informatics. Combining data from multiple modalities may help in predictive tasks. To assess the expectations of multimodal data, we introduce a comprehensive fusion framework designed to integrate temporal variables, medical images, and clinical notes in Electronic Health Record (EHR) for enhanced performance in downstream predictive tasks. Early, joint, and late fusion strategies were employed to effectively combine data from various modalities. Model performance and contribution scores show that multimodal models outperform uni-modal models in various tasks. Additionally, temporal signs contain more information than CXR images and clinical notes in three explored predictive tasks. Therefore, models integrating different data modalities can work better in predictive tasks.

Low-order fine roots of Picea asperata have different physiological mechanisms in response to seasonal freeze and freeze-thaw of soil

Seasonal soil freezing (F) and freeze-thaw cycles (FTCs) are common natural phenomena in high latitude or altitude areas of the world, and seriously affect plant physiological processes. However, studies on the effect of soil F and FTCs on fine roots are less common, especially in subalpine coniferous forests of western Sichuan, China. We set up a controlled experiment in growth chambers to explore the effects of F and FTCs on low-order fine roots of Picea asperata and differential responses of first-order roots and the first three root orders (1st, 2nd and 3rd order roots combined as a unit). Soil F and FTCs resulted in serious damage to cell membranes and root vitality of low-order fine roots, accompanied by increased MDA content and O₂ ·^- production. FTCs had a stronger effect than F treatment. In turn, low-order fine roots are the unit that responds to cold stress. These roots had increased unsaturated fatty acid contents, antioxidant enzyme activities, osmolytes and plant hormones contents when acclimation to cold stress. The first-order roots were more sensitive to cold stress than the combined first three root orders for several processes (e.g. antioxidant enzymes, osmolytes and hormones) because of their specific structure and physiological activity. This study explains physiological differences in responses of fine roots of different root orders to seasonal soil freezing, which will improve the understanding of fine root heterogeneity and support agriculture and forest management.

A fair and interpretable network for clinical risk prediction: a regularized multi-view multi-task learning approach

In healthcare domain, complication risk profiling which can be seen as multiple clinical risk prediction tasks is challenging due to the complex interaction between heterogeneous clinical entities. With the availability of real-world data, many deep learning methods are proposed for complication risk profiling. However, the existing methods face three open challenges. First, they leverage clinical data from a single view and then lead to suboptimal models. Second, most existing methods lack an effective mechanism to interpret predictions. Third, models learned from clinical data may have inherent pre-existing biases and exhibit discrimination against certain social groups. We then propose a multi-view multi-task network (MuViTaNet) to tackle these issues. MuViTaNet complements patient representation by using a multi-view encoder to exploit more information. Moreover, it uses a multi-task learning to generate more generalized representations using both labeled and unlabeled datasets. Last, a fairness variant (F-MuViTaNet) is proposed to mitigate the unfairness issues and promote healthcare equity. The experiments show that MuViTaNet outperforms existing methods for cardiac complication profiling. Its architecture also provides an effective mechanism for interpreting the predictions, which helps clinicians discover the underlying mechanism triggering the complication onsets. F-MuViTaNet can also effectively mitigate the unfairness with only negligible impact on accuracy.

Predicting Age-Related Macular Degeneration Progression with Contrastive Attention and Time-Aware LSTM

Age-related macular degeneration (AMD) is the leading cause of irreversible blindness in developed countries. Identifying patients at high risk of progression to late AMD, the sight-threatening stage, is critical for clinical actions, including medical interventions and timely monitoring. Recently, deep-learning-based models have been developed and achieved superior performance for late AMD prediction. However, most existing methods are limited to the color fundus photography (CFP) from the last ophthalmic visit and do not include the longitudinal CFP history and AMD progression during the previous years' visits. Patients in different AMD subphenotypes might have various speeds of progression in different stages of AMD disease. Capturing the progression information during the previous years' visits might be useful for the prediction of AMD progression. In this work, we propose a Contrastive-Attention-based Time-aware Long Short-Term Memory network (CAT-LSTM) to predict AMD progression. First, we adopt a convolutional neural network (CNN) model with a contrastive attention module (CA) to extract abnormal features from CFPs. Then we utilize a time-aware LSTM (T-LSTM) to model the patients' history and consider the AMD progression information. The combination of disease progression, genotype information, demographics, and CFP features are sent to T-LSTM. Moreover, we leverage an auto-encoder to represent temporal CFP sequences as fixed-size vectors and adopt k-means to cluster them into subphenotypes. We evaluate the proposed model based on real-world datasets, and the results show that the proposed model could achieve 0.925 on area under the receiver operating characteristic (AUROC) for 5-year late-AMD prediction and outperforms the state-of-the-art methods by more than 3%, which demonstrates the effectiveness of the proposed CAT-LSTM. After analyzing patient representation learned by an auto-encoder, we identify 3 novel subphenotypes of AMD patients with different characteristics and progression rates to late AMD, paving the way for improved personalization of AMD management. The code of CAT-LSTM can be found at GitHub.

Deconfounding Actor-Critic Network with Policy Adaptation for Dynamic Treatment Regimes

Despite intense efforts in basic and clinical research, an individualized ventilation strategy for critically ill patients remains a major challenge. Recently, dynamic treatment regime (DTR) with reinforcement learning (RL) on electronic health records (EHR) has attracted interest from both the healthcare industry and machine learning research community. However, most learned DTR policies might be biased due to the existence of confounders. Although some treatment actions non-survivors received may be helpful, if confounders cause the mortality, the training of RL models guided by long-term outcomes (e.g., 90-day mortality) would punish those treatment actions causing the learned DTR policies to be suboptimal. In this study, we develop a new deconfounding actor-critic network (DAC) to learn optimal DTR policies for patients. To alleviate confounding issues, we incorporate a patient resampling module and a confounding balance module into our actor-critic framework. To avoid punishing the effective treatment actions non-survivors received, we design a short-term reward to capture patients' immediate health state changes. Combining short-term with long-term rewards could further improve the model performance. Moreover, we introduce a policy adaptation method to successfully transfer the learned model to new-source small-scale datasets. The experimental results on one semi-synthetic and two different real-world datasets show the proposed model outperforms the state-of-the-art models. The proposed model provides individualized treatment decisions for mechanical ventilation that could improve patient outcomes.

First live birth in China after cryopreserved ovarian tissue transplantation to prevent premature ovarian insufficiency

OBJECTIVE

This article reports the first live birth after cryopreserved ovarian tissue transplantation to prevent premature ovarian insufficiency in China.

METHODS

A patient with myelodysplastic syndrome received ovarian tissue cryopreservation before hematopoietic stem cell transplantation, and six ovarian cortex strips were thawed and transplanted into her peritoneal pocket 2 years later.

RESULTS

Pregnancy occurred spontaneously 27 months after grafting, and a healthy girl was born at 38 weeks gestation. Until now, the child has developed normally without any major diseases.

CONCLUSIONS

We report the first live birth resulting from ovarian tissue cryopreservation and transplantation in China.

LncRNA Malat-1 From MSCs-Derived Extracellular Vesicles Suppresses Inflammation and Cartilage Degradation in Osteoarthritis

Purpose: Extracellular Vesicles (EVs) derived from hMSCs, have the potential to alleviate cartilage damage and inflammation. We aimed to explore the effects of EVs derived from lncRNA malat‐1-overexpressing human mesenchymal stem cells (hMSCs) on chondrocytes.

Material and Methods: hMSCs-derived Extracellular Vesicles (hMSCs-EVs) were identified by transmission electron microscopy and western blot. We used a Sprague-Dawley (SD) rat model of CollagenaseⅡ-induced osteoarthritis (OA) as well as IL-1β-induced OA chondrocytes. Lentiviral vectors were used to overexpress lncRNA malat‐1 in hMSCs. Chondrocyte proliferation, inflammation, extracellular matrix degradation, and cell migration were measured by Edu staining, ELISA, western blot analysis, and transwell assay. Chondrocyte apoptosis was evaluated by flow cytometry, Hoechst 33342/PI Staining, and western blot. Safranine O-fast green (S-O) staining and HE staining were used to assess morphologic alterations of the rat knee joint.

Results: hMSCs^malat−1-EVs decreased MMP-13, IL-6, and Caspase-3 expression in IL-1β-induced OA chondrocytes. Moreover, hMSCs^malat−1-EVs promoted chondrocyte proliferation and migration, suppressed apoptosis, and attenuated IL-1β-induced chondrocyte injury. Our animal experiments suggested that hMSCs^malat−1-EVs were sufficient to prevent cartilage degeneration.

Conclusion: Our findings show that lncRNA malat-1from hMSCs‐delivered EVs can promote chondrocyte proliferation, alleviate chondrocyte inflammation and cartilage degeneration, and enhance chondrocyte repair. Overall, hMSCs^malat−1-EVs might be a new potential therapeutic option for patients with OA.

[MiR-671-5p negatively regulates SMAD3 to inhibit migration and invasion of osteosarcoma cells]

OBJECTIVE

To explore the role of miR-671-5p in regulating the migration and invasion of osteosarcoma and the underlying mechanisms.

METHODS

The differentially expressed microRNAs (miRNAs) in osteosarcoma were screened in the NCBI online database, and the target proteins of these miRNAs were predicted and their functions were analyzed. Osteosarcoma cells were transfected with a plasmid overexpressing miR-671-5p, and the transfection efficiency was assessed using quantitative real-time PCR (qRT-PCR). The changes in the migration and invasion of the transfected cells were examined with Transwell assay, and the expressions of proteins related with epithelial-mesenchymal transition (EMT) were detected using Western blotting. Dual-luciferase reporter assay was performed to determine whether the 3'UTR of SMAD3 contained a targeted binding site of miR-671-5p.

RESULTS

MiR-671-5p was significantly down-regulated in both osteosarcoma tissues and osteosarcoma cells (P < 0.05). The osteosarcoma cells overexpressing miR-671-5p showed significantly reduced migration and invasion abilities (P < 0.05) with obviously lowered expressions of EMT-related proteins (P < 0.05). SMAD3 was highly expressed in osteosarcoma cells (P < 0.05), and dual-luciferase reporter assay confirmed the presence of a targeted binding site between miR-671-5p and the 3'UTR of SMAD3 (P < 0.05). In osteosarcoma cells transfected with a SMAD3-overexpressing plasmid (P < 0.05), the high expression of SMAD3 significantly inhibited by miR-671-5p overexpression (P < 0.05). Transwell assay demonstrated that SMAD3 overexpression significantly promoted the migration and invasion of osteosarcoma cells (P < 0.05), and while miR-671-5p overexpression obviously reversed this effect (P < 0.05).

CONCLUSION

MiR-671-5p can inhibit the invasion and migration of osteosarcoma cells by negatively regulating SMAD3.

An in vivo model of postinflammatory hyperpigmentation and erythema: clinical, colorimetric and molecular characteristics

BACKGROUND

Postinflammatory hyperpigmentation (PIH) is a common, acquired pigmentary disorder of the skin associated with significant quality-of-life impairment, especially in individuals with skin of colour. Current treatment for PIH is limited, largely due to a poor understanding of disease pathogenesis and the lack of a representative disease model.

OBJECTIVES

This study is intended to further develop, update and validate our previously designed in vivo model of acne-induced PIH/postinflammatory erythema (PIE) using different concentrations of trichloroacetic acid (TCA), a medium-depth chemical peel.

METHODS

Twenty-nine patients with skin types II-VI and clinician-confirmed presence of two or more truncal acne pustules and PIH/PIE were included. On the basis of Investigator's Global Assessment (IGA), clinical polarized photography (CPP), colorimetry and Skindex, we experimentally determined an optimum TCA concentration and assessed our model's ability to exhibit a dose-response relationship between degree of inciting insult and severity of resulting pigmentation. We also performed differential microRNA profiling and pathway analysis to explore the potential of microRNAs as molecular adjuncts to our model.

RESULTS

Application of TCA 30% produced lesions indistinguishable from acne-induced PIH and PIE lesions on the basis of colorimetry data without causing epidermal necrosis. Application of progressively increasing TCA doses from 20% to 30% resulted in concentration-dependent increases in CPP, IGA and colorimetry scores at all timepoints during the study. miRNA-31 and miRNA-23b may play a role in PIH pathogenesis, although further validation is required.

CONCLUSIONS

Our TCA-based in vivo model, using TCA concentrations between 20% and 30% with an optimum of 30%, enables the quantitative assessment of the pigmentary response to varying degrees of cutaneous inflammation in a fashion that mirrors natural acne-induced PIH and PIE.

Antibiotic resistance and molecular characteristics of Staphylococcus aureus isolated from pigs in Hunan, China

Staphylococcus aureus (S. aureus) has been recognized as one of the important zoonotic pathogens. However, it was limited about the epidemiology and genetic characteristics of S. aureus isolated from pigs in Hunan province, china. The aim of this study was to determine the characteristics of 163 S. aureus isolated from 590 pigs in Hunan Province, China. All isolates were characterized by agr typing, detection of virulence genes and antibiotic resistance genes, lethal test of mice and antibiotic susceptibility tests. The results showed that 30 strains of the 163 isolates were divided into agrⅠ (18.40%), agrⅡ(36/163, 22.09%), agrⅢ (20/163, 12.27%,), agr Ⅳ(20/163,12.27%) and the remaining 57 isolates were amplified negative by agr primers. In the 163 isolates, the detection rate of the virulence genes hlb, hld, hla, icaA, seb, fnbA, eta, etb, sea, tst and pvl ranged from 2.45% to 100%. The 43 isolates that were lethal to the mice, had β-hemolytic activity, the number of virulence genes of which was 7.8% higher than that of the remaining 120 non-fatal strains. The resistance rates of the 163 isolates to the 15 antibiotics were 0% (0/163) - 100% (163/163). All isolates were susceptible to Vancomycin and only 7 isolates were methicillin - resistant S. aureus (MRSA). The detection rates of the 11 resistance genes was 0% (0/163) - 100% (163/163). This study first to describes the epidemiology and characteristics of S. aureus from pigs in Hunan Province, which will help in tracking the evolution of epidemic strains and preventing pig-human transmission events.

First pregnancy in China after ovarian tissue transplantation to prevent premature ovarian insufficiency

OBJECTIVE

This article reports the first case of pregnancy after frozen-thawed ovarian tissue transplantation to prevent iatrogenic premature ovarian insufficiency in China.

METHODS

Ovarian tissue cryopreservation was performed in a patient with myelodysplastic syndrome (MDS) before multi-agent chemotherapy and hematopoietic stem cell transplantation. Two years later, she showed complete remission from MDS, and six frozen-thawed ovarian tissue strips were transplanted into the peritoneal pocket.

RESULTS

The patient's ovarian activity was restored 3 months after transplantation, and pregnancy occurred spontaneously 27 months after grafting. Until now, the pregnancy has progressed for 30 weeks, and the repeated ultrasound showed normal fetal development.

CONCLUSION

This is the first pregnancy resulting from ovarian tissue cryopreservation and transplantation in China.

Application of an individualized nomogram in first-trimester screening for trisomy 21

OBJECTIVES

To develop and validate a nomogram based on fetal nuchal translucency thickness (NT) and ultrasonographic facial markers for screening for trisomy 21 in the first trimester of pregnancy.

METHODS

This was a retrospective case-control study using stored two-dimensional midsagittal fetal profile images captured at 11 + 0 to 13 + 6 weeks' gestation in singleton pregnancies. We included images from 302 trisomy-21 pregnancies and 322 euploid pregnancies. Cases were divided into a training set (200 euploid + 200 with trisomy 21) and a validation set (122 euploid + 102 with trisomy 21) at a ratio of approximately 2:1. For each, the maternal age, gestational age, fetal NT and karyotype were noted, and 12 ultrasonographic fetal facial markers were measured. The least absolute shrinkage and selection operator (LASSO) method and multivariable analysis were used to select automatically the discriminative markers. Logistic regression was used to develop a LASSO model, based on the selected markers, to screen for trisomy 21 in the first trimester of pregnancy. Furthermore, 60 of the 624 images were selected randomly as a retest set to evaluate the model's robustness. The predictive performance of screening for trisomy 21 of a model based on fetal NT and maternal age and of the LASSO model was assessed using the area under the receiver-operating-characteristics curve (AUC). A nomogram was developed as an individualized tool to predict patient-specific probability for trisomy 21, which is a more visual presentation of the LASSO model. The performance of the nomogram was assessed using the C-index and calibration curve.

RESULTS

Into the LASSO model were incorporated eight markers, including fetal NT, prenasal-thickness-to-nasal-bone-length ratio, facial profile line, frontomaxillary facial angle, frontonasal facial angle, mandibulomaxillary facial angle, maxilla-nasion-mandible angle and d2 (distance between the anterior edge of the prefrontal skin and the mandibulomaxillary line) (all P < 0.05). The AUCs of the LASSO model for screening for trisomy 21 were 0.983 (95% CI, 0.971-0.994) in the training set and 0.979 (95% CI, 0.966-0.993) in the validation set, and these were higher than the AUCs of all eight individual ultrasonographic markers included in the model. The AUC of the LASSO model in the retest set was 0.997 (95% CI, 0.990-1.000), indicating good robustness of the LASSO model. The AUC of the LASSO model was significantly higher than that of the model based on fetal NT and maternal age in both training and validation sets (P < 0.001 for both). The nomogram of the LASSO model showed good discrimination of trisomy 21, with C-indices of 0.983 in the training set and 0.981 in the validation set.

CONCLUSIONS

We present an individualized nomogram which incorporates fetal NT and a series of ultrasonographic facial profile markers selected by the LASSO method and multivariable analysis. This nomogram can potentially be utilized as a convenient and effective tool in screening for trisomy 21 in the first trimester of pregnancy. © 2020 The Authors. Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology.

A survival model generalized to regression learning algorithms

Survival prediction is an important problem that is encountered widely in industry and medicine. Despite the explosion of artificial intelligence technologies, no uniformed method allows the application of any type of regression learning algorithm to a survival prediction problem. Here, we present a statistical modeling method that is generalized to all types of regression learning algorithm, including deep learning. We present its empirical advantage when it is applied to traditional survival problems. We demonstrate its expanded applications in different types of regression learning algorithm, such as gradient boosted trees, convolutional neural networks and recurrent neural networks. Additionally, we demonstrate its application in clinical informatic data, pathological images and the hardware industry. We expect that this algorithm will be widely applicable for diverse types of survival data, including discrete data types and those suitable for deep learning such as those with time or spatial continuity.

Brain Atlas Guided Attention U-Net for White Matter Hyperintensity Segmentation

White Matter Hyperintensities (WMH) are the most common manifestation of cerebral small vessel disease (cSVD) on the brain MRI. Accurate WMH segmentation algorithms are important to determine cSVD burden and its clinical con-sequences. Most of existing WMH segmentation algorithms require both fluid attenuated inversion recovery (FLAIR) images and T1-weighted images as inputs. However, T1-weighted images are typically not part of standard clinical scans which are acquired for patients with acute stroke. In this paper, we propose a novel brain atlas guided attention U-Net (BAGAU-Net) that leverages only FLAIR images with a spatially-registered white matter (WM) brain atlas to yield competitive WMH segmentation performance. Specifically, we designed a dual-path segmentation model with two novel connecting mechanisms, namely multi-input attention module (MAM) and attention fusion module (AFM) to fuse the information from two paths for accurate results. Experiments on two publicly available datasets show the effectiveness of the proposed BAGAU-Net. With only FLAIR images and WM brain atlas, BAGAU-Net outperforms the state-of-the-art method with T1-weighted images, paving the way for effective development of WMH segmentation. Availability: https://github.com/Ericzhang1/BAGAU-Net.

[A nested case-control study on relationship of traditional and combined lipid metabolism indexes with incidence of diabetes]

Objective: To explore the relationship between lipid indicators and the incidence of diabetes, and to compare the diabetes prediction and identification power of traditional lipid combined lipid indicators, in order to explore the best alternative indicators for identifying and predicting diabetes. Methods: Based on the Jinchang cohort, a nested case-control study was conducted in 1 025 new cases of diabetes after excluding patients with malignant tumor and related endocrine, circulatory system disease, then an age (±2 years), gender matched 1∶1 control group of 1 025 cases was set to analyze the relationship between the incidence of diabetes and lipid parameters. Results: Among the traditional lipid parameters, the fourth quartile of TG, TC, and LDL-C indicated higher risks of developing diabetes, which was 14.00 times (95%CI: 9.73-20.15), 2.15 times (95%CI: 1.65-2.79) and 1.66 times (95%CI: 1.29-2.14) than that of the first quartile, respectively. The risk of developing diabetes indicated by the fourth quartile of HDL-C was 0.21 times than that indicated by the first quartile (95%CI: 0.15-0.28). In the combined lipid parameters, the fourth quartile of TG/HDL-C, TC/HDL-C, LDL-C/HDL-C and non-HDL-C indicated higher risks of developing diabetes, which was 14.86 times (95%CI: 10.35-21.34), 8.12 times (95%CI: 5.94-11.01), 5.85 times (95%CI:4.34-7.88) and 5.20 times (95%CI: 3.85-7.03) than that indicated by the first quartile, respectively. The areas under the ROC curve of TG, TC, HDL-C, LDL-C, TG/HDL-C, TC/HDL-C, LDL-C/HDL-C and non-HDL-C were 0.76 (95%CI: 0.74-0.78), 0.59 (95%CI: 0.57-0.61), 0.67 (95%CI: 0.65-0.69), 0.57 (95%CI: 0.55-0.59), 0.77 (95%CI: 0.75-0.78), 0.73 (95%CI: 0.71-0.75), 0.69 (95%CI: 0.67-0.71) and 0.66 (95%CI: 0.64-0.68), respectively. The optimal diabetes predicting point cuts of TG, TC, HDL-C, LDL-C, TG/HDL-C, TC/HDL-C, LDL-C/HDL-C and non-HDL-C were 1.40, 4.70, 1.28, 3.25, 1.17, 3.43, 2.46, and 3.58 mmol/L, respectively. Conclusions: Lipid metabolic disorder is a risk factor for diabetes. TG and TG/HDL-C are the good lipid metabolism indicators for the prediction of diabetic.

[Relationship of body mass index and blood pressure with diabetes: a nested case-control study]

Objective: To explore the relationship of body mass index and blood pressure with the incidence of diabetes in Jinchang cohort. Methods: We designed a nested case-control study, a total of 29 572 workers who had no history of diabetes in baseline survey in Jinchang cohort were selected as the study cohort from June 2011 to December 2013. After 2 year follow-up, 1 021 workers with first diagnosed diabetes were selected as the case group, after 1∶1 matching according to the same gender and age ±2 years among those without diabetes, circulatory system, or endocrine system diseases during the same follow-up period, 1 021 controls was selected and 2 042 subjects were finally included. We used multivariate conditional logistic regression model, additive interaction model and multiplicative interaction model to explore the relationship of body mass index and blood pressure with the incidence of diabetes. Results: After adjusting for factors such as occupation, alcohol use, family history of diabetes, hyperuricemia, hypercholesterolemia, hypertriglyceridemia, low-HDL cholesterolemia and high-LDL cholesterolemia, multivariate conditional logistic regression analysis showed that the risk of diabetes increased with body mass index and blood pressure. Hypertension and overweight/obesity had a multiplicative interaction on the incidence of diabetes. The risks of diabetes in men and women with hypertension and overweight/obese were 2.04 times (95%CI: 1.54-2.69) and 3.88 times (95%CI: 2.55-5.91) higher than those in men and women with normal body weight and blood pressure, respectively. In the combination of BMI and blood pressure, obese individuals with SBP≥160 mmHg were 4.57 times (95%CI: 2.50-8.34) more likely to have diabetes than those with normal BMI and SBP, obese individuals with DBP≥90 mmHg were 3.40 times (95%CI: 2.19-5.28) more likely to have diabetes than those with normal BMI and DBP. Conclusions: Overweight/obesity and hypertension can increase the risk of diabetes. Health education about body weight and blood pressure controls should be strengthened to reduce the risk of diabetes.

[Effect of HBV infection pattern on prevalence of fatty liver disease in Jinchang cohort]

Objective: To investigate the influence of HBV infection on the prevalence of fatty liver disease in Jinchang cohort and provide theoretical evidence for the prevention and treatment of fatty liver disease. Methods: Epidemiological investigation, laboratory examination and abdominal ultrasound were conducted in the baseline population of Jinchang cohort to collect the basic data, the differences in the prevalence of fatty liver disease under different HBV infection patterns were described and compared and the influence of different HBV infection patterns on the prevalence of fatty liver disease were evaluated by using logistic regression analysis. Results: The baseline Jinchang cohort population totaled 45 605, including 27 917 males and 17 688 females. The male to female ratio was 1.6∶1. The mean age of the overall population was 46.49 years. Among the 8 common HBV infection modes in the Jinchang cohort, the prevalence of fatty liver was low in HBsAg, HBeAg and HBcAb positive, HBsAg and HBcAb positive, and HBsAg, HBeAb and HBcAb positive groups. For 4 serum markers of HBV infection, the prevalence of fatty liver disease in HBsAg and HBeAg positive groups was lower than that in HBsAg and HBeAg negative groups. Logistic regression analysis showed that being HBsAg and HBcAb positive (OR=0.61, 95%CI: 0.39-0.98) and HBsAg, HBeAg and HBcAb positive (OR=0.52, 95%CI: 0.30-0.89) could reduce the risk for fatty liver disease. Conclusion: Acute HBV infection reduces the prevalence of fatty liver disease, and the reason may be related to the disturbance of the body's fat metabolism by active HBV replication.

[Prospective cohort study of relationship of triglyceride, fasting blood-glucose and triglyceride glucose product index with risk of hypertension]

Objective: To investigate the relationship of triglyceride (TG), fasting blood glucose (FPG) and triglyceride glucose product index (TyG) with the incidence of hypertension, and provide basic data for the prevention and treatment of hypertension in the population. Methods: A total of 23 581 individuals who met the research criteria in Jinchang cohort were selected as the research subjects, the Cox proportional hazard model was used to analyze the relationship of TG, FPG, and TyG with the risk of hypertension. A stratified analysis was conducted by sex. Results: After adjusting for confounding factors, compared with the normal TG group, the HR(95%CI) of the elevated TG margin group and the elevated group were 1.16 (1.01-1.34) and 1.49 (1.30-1.70), respectively in the total population. Among men, they were 1.13 (1.01-1.27) and 1.17 (1.06-1.30), and among women, they were 1.05 (0.88-1.26) and 1.06 (0.88-1.28). Compared with the normal FPG group, the HR (95%CI) of the FPG-impaired group were 1.29 (1.13-1.48) in the total population, 1.26 (1.08-1.48) in men and 1.59 (1.14-2.21) in women. Taking the lowest quartile array as a reference, the HR (95%CI) of the highest quartile array of TyG was 1.73 (1.45-2.07) in the total population, 1.32 (1.14-1.53) in men and 1.87 (1.37-2.54) in women. TG, FPG had a nonlinear dose-response relationship with the risk of hypertension, while TyG had a linear correlation with the risk of hypertension. Conclusions: Higher TG, FPG, and TyG levels are independent risk factors for the incidence of hypertension. People with higher TG, FPG and TyG are at high risk for hypertension, to which close attention should be paid in the prevention and treatment of hypertension.

[Analysis on influencing factors for nonalcoholic fatty liver disease in Jinchang cohort]

Objective: To explore the influencing factors for non-alcoholic fatty liver disease (NAFLD) in Jinchang cohort, and provide scientific basis for the prevention and control of NAFLD. Methods: A total of 20 051 patients without fatty liver at baseline survey and met the inclusion criteria in Jinchang cohort were selected as study subjects. Prospective cohort study and Cox regression analysis were used to investigate the influencing factors for NAFLD, and the dose-response relationship between related biochemical indicators and NAFLD risk was studied by restricted cubic spline method. Results: The incidence of NAFLD was 42.37/1 000 person years. Multivariate Cox regression analysis showed that being worker and technical personnel (being worker:HR=0.84,95%CI:0.70-0.99;being technical personnel:HR=0.73,95%CI:0.56-0.95), tea drinking (current drinking:HR=0.86,95%CI:0.78-0.94;previous drinking: HR=0.52,95%CI: 0.31-0.86), exercise (occasionally: HR=0.79, 95%CI: 0.68-0.91;frequently:HR=0.60,95%CI:0.52-0.69), low body weight (HR=0.10, 95%CI: 0.05-0.22), daily intake of dairy products >300 ml/day (HR=0.78, 95%CI: 0.71-0.87) and HBV infection (HR=0.77, 95%CI: 0.60-0.99) were the protective factors for NAFLD, while being internal or office workers (HR=1.84, 95%CI: 1.46-2.31), income ≥2 000 yuan (2 000- yuan: HR=1.32, 95%CI: 1.04-1.66; ≥5 000 yuan: HR=1.72, 95%CI:1.11-2.66), bachelor degree or above (HR=1.35,95%CI:1.03-1.76), overweight (HR=2.31, 95%CI:2.08-2.55), obesity (HR=3.95, 95%CI: 3.42-4.56), impaired fasting blood glucose (HR=1.31, 95%CI:1.17-1.47), diabetes (HR=1.53, 95%CI: 1.30-1.80), increased TC (HR=1.37,95%CI:1.24-1.52), increased TG (HR=1.79,95%CI: 1.62-1.98), decreased HDL-C (HR=1.29, 95%CI: 1.14-1.45), increased ALT (HR=1.13, 95%CI: 1.01-1.26) and high-fat diet (HR=1.24, 95%CI: 1.11-1.40) were the risk factors for NAFLD. Moreover, TC, TG, HDL-C, ALT and FPG all showed good dose-response relationship with the incidence of NAFLD. Conclusion: Occupation, education level, income level, tea drinking, exercise, BMI, FPG, blood lipid, ALT, HBV infection and diet were related to the incidence of NAFLD.

[MiR-300 inhibits invasion and metastasis of osteosarcoma cell MG63 by negatively regulating PTTG1]

OBJECTIVE

To investigate the effects of miR-300 and PTTG1 on osteosarcoma invasion and metastasis and explore the molecular mechanism of osteosarcoma invasion and metastasis.

OBJECTIVE

Western blot was used to detect the expression of PTTG1 in human osteoblasts hFOB1.19 and osteosarcoma cell MG63 and to detect the transfection efficiency of cells transfected with PTTG1-knockdown plasmid; Transwell invasion assay and CCK8 assay detected the effects of knockdown of PTTG1 and overexpression of miR-300 on the invasion and proliferation of osteosarcoma cell MG63. On-line prediction and screening of microRNAs (miRNAs) with complementary PTTG1 binding was conducted. qRT-PCR was performed to examine the expression of miR-300 in hFOB1.19 and MG63 cells, and Western blotting was used to detect the expression of PTTG1 in MG63 cells after transfection with a miR- 300 plasmid. Double luciferase assay was used to detect the targeted binding of miR-300 and PTTG, Transwell invasion assay and CCK8 assay were used to detect the effects of overexpression of miR-300 and overexpression of PTTG1 plasmid on invasion and proliferation of osteosarcoma cell line MG63.

OBJECTIVE

PTTG1 was highly expressed in MG63 cells (P=0.0002). PTTG1 knockdown significantly inhibited the invasion (P=0.0002) and proliferation (P=0.0039) of MG63 cells. Based on the results of online prediction of complementary miRNAs to PTTG1 and analysis of the data from NCBI database, miR-300 was determined as the target miRNA in this study. qRT-PCR results showed a significantly decreased expression of miR-300 in MG63 cells (P=0.0004). Overexpression of MiR-300 in MG63 cells significantly decreased the expression of PTTG1 (P=0.0007), and the expressions of miR-300 and PTTG1 were negatively correlated. Dual luciferase assay showed that miR-300 could specifically bind to PTTG1 (P=0.001). Overexpression of PTTG1 could significantly reverse the effect of miR-300 overexpression on invasion (P=0.0003) and proliferation (P=0.0077) of MG63 cells.

OBJECTIVE

Overexpression of miR-300 can inhibit the invasion and metastasis of osteosarcoma cell MG63 by targeting PTTG1.

An interpretable deep-learning model for early prediction of sepsis in the emergency department

Sepsis is a life-threatening condition with high mortality rates and expensive treatment costs. Early prediction of sepsis improves survival in septic patients. In this paper, we report our top-performing method in the 2019 DII National Data Science Challenge to predict onset of sepsis 4 h before its diagnosis on electronic health records of over 100,000 unique patients in emergency departments. A long short-term memory (LSTM)-based model with event embedding and time encoding is leveraged to model clinical time series and boost prediction performance. Attention mechanism and global max pooling techniques are utilized to enable interpretation for the deep-learning model. Our model achieved an average area under the curve of 0.892 and was selected as one of the winners of the challenge for both prediction accuracy and clinical interpretability. This study paves the way for future intelligent clinical decision support, helping to deliver early, life-saving care to the bedside of septic patients.

The advanced lung cancer inflammation index predicts outcomes in patients with Crohn's disease after surgical resection

AIM

Precise biomarkers for predicting prognosis could help to identify high-risk Crohn's disease (CD) patients to facilitate better follow-up during the postoperative course. In this study, the primary aim is the identification of the most reliable nutrition marker that predicts surgical relapse in CD patients.

METHOD

We first evaluated the predictive value of various nutrition markers for postoperative surgical relapse in CD patients and identified the advanced lung cancer inflammation index (ALI) as a promising biomarker. Then, we assessed the clinical significance of preoperative ALI in CD patients using two cohorts.

RESULTS

Preoperative ALI showed the highest correlation with reoperation rate compared with other nutritional parameters in CD patients receiving surgical resection (sensitivity 53%, specificity 86%, area under the curve 0.71). Lower levels of preoperative ALI were significantly correlated with the presence of perianal disease. A lower level of preoperative ALI was an independent prognostic factor for reoperation rate after an intestinal resection (hazard ratio 3.37, 95% CI 1.38-10.12, P = 0.006), and the prognostic impact of preoperative ALI was successfully validated in an independent cohort using the same cut-off value.

CONCLUSION

Preoperative ALI might be useful for postoperative management of CD patients.

An interpretable risk prediction model for healthcare with pattern attention

BACKGROUND

The availability of massive amount of data enables the possibility of clinical predictive tasks. Deep learning methods have achieved promising performance on the tasks. However, most existing methods suffer from three limitations: (1) There are lots of missing value for real value events, many methods impute the missing value and then train their models based on the imputed values, which may introduce imputation bias. The models' performance is highly dependent on the imputation accuracy. (2) Lots of existing studies just take Boolean value medical events (e.g. diagnosis code) as inputs, but ignore real value medical events (e.g., lab tests and vital signs), which are more important for acute disease (e.g., sepsis) and mortality prediction. (3) Existing interpretable models can illustrate which medical events are conducive to the output results, but are not able to give contributions of patterns among medical events.

METHODS

In this study, we propose a novel interpretable Pattern Attention model with Value Embedding (PAVE) to predict the risks of certain diseases. PAVE takes the embedding of various medical events, their values and the corresponding occurring time as inputs, leverage self-attention mechanism to attend to meaningful patterns among medical events for risk prediction tasks. Because only the observed values are embedded into vectors, we don't need to impute the missing values and thus avoids the imputations bias. Moreover, the self-attention mechanism is helpful for the model interpretability, which means the proposed model can output which patterns cause high risks.

RESULTS

We conduct sepsis onset prediction and mortality prediction experiments on a publicly available dataset MIMIC-III and our proprietary EHR dataset. The experimental results show that PAVE outperforms existing models. Moreover, by analyzing the self-attention weights, our model outputs meaningful medical event patterns related to mortality.

CONCLUSIONS

PAVE learns effective medical event representation by incorporating the values and occurring time, which can improve the risk prediction performance. Moreover, the presented self-attention mechanism can not only capture patients' health state information, but also output the contributions of various medical event patterns, which pave the way for interpretable clinical risk predictions.

AVAILABILITY

The code for this paper is available at: https://github.com/yinchangchang/PAVE .

High expression of SHMT2 is correlated with tumor progression and predicts poor prognosis in gastrointestinal tumors

OBJECTIVE

Gastrointestinal tumors are malignant tumors with high morbidity. Mitochondrial serine hydroxymethyltransferase 2 (SHMT2) is a key enzyme in the synthesis of serine and glycine, which has prognostic and therapeutic value for many malignant tumors. However, the role of SHMT2 in gastric cancer (GC), esophageal cancer (ESCC), and colorectal cancer (CC) has not been clarified.

PATIENTS AND METHODS

The expression of SHMT2 was detected in GC, ESCC, and CC by immunohistochemistry and reverse real time transcription-polymerase chain reaction. The relationships between SHMT2 expression and clinicopathologic characteristics, recurrence-free survival (RFS), and disease-specific survival (DSS) were analyzed by the survival analysis and correlation analysis.

RESULTS

The positive expression rate of SHMT2 in GC, ESCC, and CC was 74.1%, 69.2%, and 71.7%, respectively. Patients with high expression of SHMT2 had a worse prognosis. In GC, high SHMT2 expression had positive correlation with lymph node metastasis (p=0.005) and histological grade (p=0.002). In ESCC, high SHMT2 expression had positive correlation with pT classification (p=0.033) and pM classification (p=0.029). In CC, high SHMT2 expression had positive correlation with tumor size (p=0.004), lymph node metastasis (p=0.035), TNM stage (p=0.007), and histological grade (p=0.020). Notably, SHMT2 expression was an independent prognostic factor for RFS and DSS in GC, ESCC, and CC (p<0.05).

CONCLUSIONS

SHMT2 is upregulated in GC, ESCC, and CC. The high expression of SHMT2 is correlated with gastrointestinal tumors progression, and poor prognosis, which is a potential new target for the diagnosis and treatment of gastrointestinal tumors.

Combining structured and unstructured data for predictive models: a deep learning approach

Background

The broad adoption of electronic health records (EHRs) provides great opportunities to conduct health care research and solve various clinical problems in medicine. With recent advances and success, methods based on machine learning and deep learning have become increasingly popular in medical informatics. However, while many research studies utilize temporal structured data on predictive modeling, they typically neglect potentially valuable information in unstructured clinical notes. Integrating heterogeneous data types across EHRs through deep learning techniques may help improve the performance of prediction models.

Methods

In this research, we proposed 2 general-purpose multi-modal neural network architectures to enhance patient representation learning by combining sequential unstructured notes with structured data. The proposed fusion models leverage document embeddings for the representation of long clinical note documents and either convolutional neural network or long short-term memory networks to model the sequential clinical notes and temporal signals, and one-hot encoding for static information representation. The concatenated representation is the final patient representation which is used to make predictions.

Results

We evaluate the performance of proposed models on 3 risk prediction tasks (i.e. in-hospital mortality, 30-day hospital readmission, and long length of stay prediction) using derived data from the publicly available Medical Information Mart for Intensive Care III dataset. Our results show that by combining unstructured clinical notes with structured data, the proposed models outperform other models that utilize either unstructured notes or structured data only.

Conclusions

The proposed fusion models learn better patient representation by combining structured and unstructured data. Integrating heterogeneous data types across EHRs helps improve the performance of prediction models and reduce errors.

The bone marrow-derived mesenchymal stem cells (BMSCs) alleviate diabetic peripheral neuropathy induced by STZ via activating GSK-3β/β-catenin signaling pathway

BACKGROUND

Diabetic peripheral neuropathy, a common complication of diabetic mellitus, has brought a threaten on patients' health. The bone marrow-derived mesenchymal stem cells (BMSCs) were reported to play an important role in diverse diseases. Nevertheless, the specific function of BMSCs in diabetic peripheral neuropathy remained uncharacterized.

METHODS

A wide range of experiments including RT-qPCR, western blot, H&E staining, oxidative stress assessment, measurement of thermal sensitivity, ELISA, urine protein and CCK-8 assays were implemented to explore the function and mechanism of BMSCs in vivo and vitro.

RESULTS

The experimental results displayed that BMSCs improve STZ-induced diabetes symptoms in rats by decreasing blood glucose and urinary protein. Functionally, BMSCs ameliorate oxidative stress, painful diabetic neuropathy, neurotrophic status and angiogenesis in STZ-induced rats. Moreover, BMSCs participate in the regulation of sciatic neuro morphology in diabetic neuropathy rat model. In mechanism, BMSCs alleviate diabetic peripheral neuropathy via activating GSK-3β/β-catenin signaling pathway in rats and improve Schwann's cells viability by activating GSK-3β/β-catenin signaling pathway under high glucose.

CONCLUSIONS

We verified that BMSCs alleviate diabetic peripheral neuropathy of rats induced by STZ via activating GSK-3β/β-catenin signaling pathway, which implied a novel biomarker for diabetic peripheral neuropathy treatment.

Marrying Medical Domain Knowledge With Deep Learning on Electronic Health Records: A Deep Visual Analytics Approach

BACKGROUND

Deep learning models have attracted significant interest from health care researchers during the last few decades. There have been many studies that apply deep learning to medical applications and achieve promising results. However, there are three limitations to the existing models: (1) most clinicians are unable to interpret the results from the existing models, (2) existing models cannot incorporate complicated medical domain knowledge (eg, a disease causes another disease), and (3) most existing models lack visual exploration and interaction. Both the electronic health record (EHR) data set and the deep model results are complex and abstract, which impedes clinicians from exploring and communicating with the model directly.

OBJECTIVE

The objective of this study is to develop an interpretable and accurate risk prediction model as well as an interactive clinical prediction system to support EHR data exploration, knowledge graph demonstration, and model interpretation.

METHODS

A domain-knowledge-guided recurrent neural network (DG-RNN) model is proposed to predict clinical risks. The model takes medical event sequences as input and incorporates medical domain knowledge by attending to a subgraph of the whole medical knowledge graph. A global pooling operation and a fully connected layer are used to output the clinical outcomes. The middle results and the parameters of the fully connected layer are helpful in identifying which medical events cause clinical risks. DG-Viz is also designed to support EHR data exploration, knowledge graph demonstration, and model interpretation.

RESULTS

We conducted both risk prediction experiments and a case study on a real-world data set. A total of 554 patients with heart failure and 1662 control patients without heart failure were selected from the data set. The experimental results show that the proposed DG-RNN outperforms the state-of-the-art approaches by approximately 1.5%. The case study demonstrates how our medical physician collaborator can effectively explore the data and interpret the prediction results using DG-Viz.

CONCLUSIONS

In this study, we present DG-Viz, an interactive clinical prediction system, which brings together the power of deep learning (ie, a DG-RNN-based model) and visual analytics to predict clinical risks and visually interpret the EHR prediction results. Experimental results and a case study on heart failure risk prediction tasks demonstrate the effectiveness and usefulness of the DG-Viz system. This study will pave the way for interactive, interpretable, and accurate clinical risk predictions.

Study on automatic detection and classification of breast nodule using deep convolutional neural network system

Backgrounds

Conventional ultrasound manual scanning and artificial diagnosis approaches in breast are considered to be operator-dependence, slight slow and error-prone. In this study, we used Automated Breast Ultrasound (ABUS) machine for the scanning, and deep convolutional neural network (CNN) technology, a kind of Deep Learning (DL) algorithm, for the detection and classification of breast nodules, aiming to achieve the automatic and accurate diagnosis of breast nodules.

Methods

Two hundred and ninety-three lesions from 194 patients with definite pathological diagnosis results (117 benign and 176 malignancy) were recruited as case group. Another 70 patients without breast diseases were enrolled as control group. All the breast scans were carried out by an ABUS machine and then randomly divided into training set, verification set and test set, with a proportion of 7:1:2. In the training set, we constructed a detection model by a three-dimensionally U-shaped convolutional neural network (3D U-Net) architecture for the purpose of segment the nodules from background breast images. Processes such as residual block, attention connections, and hard mining were used to optimize the model while strategies of random cropping, flipping and rotation for data augmentation. In the test phase, the current model was compared with those in previously reported studies. In the verification set, the detection effectiveness of detection model was evaluated. In the classification phase, multiple convolutional layers and fully-connected layers were applied to set up a classification model, aiming to identify whether the nodule was malignancy.

Results

Our detection model yielded a sensitivity of 91% and 1.92 false positive subjects per automatically scanned imaging. The classification model achieved a sensitivity of 87.0%, a specificity of 88.0% and an accuracy of 87.5%.

Conclusions

Deep CNN combined with ABUS maybe a promising tool for easy detection and accurate diagnosis of breast nodule.

NGS-based profiling reveals a critical contributing role of somatic D-loop mtDNA mutations in HBV-related hepatocarcinogenesis

BACKGROUND

Somatic mutations of mitochondrial DNA (mtDNA) have been extensively identified mainly by traditional Sanger sequencing technology in various cancer types. However, low detection sensitivity of traditional methods greatly limits the comprehensive profiling of mtDNA somatic mutations in cancers, especially in hepatitis B virus (HBV)-related hepatocellular carcinoma (HCC). Moreover, the functional roles of mtDNA mutation in HBV-related hepatocarcinogenesis have not been systematically revealed.

PATIENTS AND METHODS

Next-generation sequencing (NGS) platform was applied to profile the somatic mtDNA mutations of HCC and paired paratumor (non-HCC) tissues from a large cohort of 156 HBV-HCC patients.

RESULTS

Our data revealed the common existence of mtDNA mutation in both inflammatory and cancer tissues with significantly different mutation pattern. The mutation density (mutation number/region length) of D-loop region was much higher than that of other regions in both HCC and non-HCC tissues. Unexpectedly, the average mutation number in D-loop region of HCC tissues was significantly less than that of non-HCC tissues. In contrast, the heteroplasmy level of D-loop region mutations was significantly increased in HCC tissues, implying that the D-loop mutations might be positively selected in HCC tissues. Furthermore, our results indicated that the patients with D-loop mutations had a significantly lower mtDNA copy number and were more likely to relapse. In vitro experiments demonstrated that proliferation, invasion and metastasis ability of HCC cells with D-loop region mutations were significantly higher than those without D-loop region mutations.

CONCLUSION

These results emphasize the critical contributing role of somatic mtDNA D-loop mutations in HBV-related hepatocarcinogenesis.

Different effects of total flavonoids from Arachniodes exilis on human umbilical cord mesenchymal stem cells in vitro

Traditional Chinese medicines are used in promotion of fractured bone healing and bone diseases. Some studies reported total flavonoids from plant can be used as an auxiliary source of exogenous.Use different methods to identify and verify effects of total flavonoids from Arachniodes exilis (TFAE) on human umbilical cord mesenchymal stem cells (HUCMSCs) in vitro.Concentrations of 1 and 5 μg/mL TFAE significantly increased ALPase activity in HUCMSCs compared to the other concentrations at days 3 and 7 (P < .05). RT-PCR showed that expression levels of osteogenic genes (Col1a1, OPN, Runx2 and Osx) were remarkably enhanced in HUCMSCs following treatment with different concentrations of TFAE for 9 days compared with 0 μg/mL TFAE group (control). The results showed that concentration < 5 μg/mL of TFAE induced osteogenic differentiation in HUCMSCs Alizarin red staining assays revealed that both TFAE and S1191 was significantly decreased (7.80 ± 0.66) compared with the TFAE group (16.00 ± 0.97) (P < .01). ALPase activity on days 3 and 7 was relatively lower in HUCMSCs grown in media supplemented with both S1191 and TFAE than that of in TFAE group only. The results indicated that osteogenic markers (Col1a1, OPN, Runx2 and Osx) were significantly downregulated in the TFAE + S1191 group in comparison to the control group. The expressions of Col1a and OPN in the TFAE + S1191 group decreased significantly (P < .01) by Western blotting.TFAE promotes the odonto/osteogenic differentiation of human UCMSCs via activation of ER.

Palmatine alleviates hyperalgesia by inhibiting the expression of calcitonin gene-related peptide in the trigeminal ganglion of rats with chronic constriction injury of the infraorbital nerve

Trigeminal neuralgia is one of the most common of the neuropathic pains, and it can seriously influence patients' quality of life. Calcitonin gene-related peptide (CGRP) is a type of nociceptive neurotransmitter that is expressed in neurons of the trigeminal ganglion and plays a major part in transmitting pain. The rat model of trigeminal neuralgia was established by causing a chronic constriction injury of the infraorbital nerve (CCI-ION). Male Sprague-Dawley rats (n=24) were randomly divided into a sham control group (sham, n=6), sham-treated with palmatine group (sham+palmatine, n=6), trigeminal nerve model group (TN, n=6), and trigeminal nerve treated with palmatine group (TN+palmatine, n=6). Fifteen days after the operation the mechanical response threshold was decreased in the TN group compared with the sham group. From postoperative day 7 to day 15, the mechanical response threshold in the TN+palmatine group significantly increased compared with the TN group. On postoperative day 15 the results of quantitative polymerase chain reaction (qPCR), immunohistochemical staining, and western blotting showed an obvious increase in expression of CGRP and its receptors, serum concentrations of interleukin-1β (IL-1β), and tumour necrosis factor-α (TNF-α), and phosphorylation of protein kinase C (PKC) in the trigeminal ganglia of the TN group compared with the sham group, but these increases could be down-regulated by treatment with palmatine. Palmatine might therefore have therapeutic potential for the treatment of trigeminal neuralgia by inhibiting the expression of CGRP and its receptors in trigeminal ganglia, suppressing the serum concentrations of IL-1β and TNF-α, and decreasing the phosphorylation of PKC in the trigeminal ganglia of affected rats.

Silencing of estrogen receptor β promotes the invasion and migration of osteosarcoma cells through activating Wnt signaling pathway

Purpose

This study aimed to evaluate the specific roles of estrogen receptor β (ERβ) on the invasion and migration of osteosarcoma (OS) cells and explore the regulatory mechanisms relating with Wnt signaling pathway.

Methods

The expression of ERβ was detected in human OS tissues by quantitative real-time PCR and immunohistochemistry. U2-OS cells were transfected with siRNA-ERβ (si-ERβ) to downregulate ERβ and treated with FH535 to inhibit Wnt signaling. The migration and invasion ability was detected by scratch and transwell assay, respectively. The expression of β-catenin, MMP-7, and MMP-9 was detected by Western blot. Subcutaneous tumor-bearing model was established by injection of U2-OS cells into mice, and the tumor volumes were measured. Orthotopic transplantation model was established by transplantation of tumor tissues into the liver of mice, and the metastatic tumors were counted.

Results

ERβ was downregulated in human OS tissues and U2-OS cells. The transfection of si-ERβ significantly increased the scratch healing rate; the number of invasion cells; and the expression of β-catenin, MMP-7, and MMP-9 in U2-OS cells. The injection of si-ERβ-transfected U2-OS cells into mice significantly increased the subcutaneous tumor volume; the expression of β-catenin, MMP-7, and MMP-9; and the number of metastatic tumors in liver tissues. The promoting effects of si-ERβ on the invasion and migration of U2-OS cells were significantly reversed by FH535 in vitro and in vivo.

Conclusion

Silencing of ERβ promotes the invasion and migration of OS cells via activating Wnt signaling pathway.

Effects of aging, baseline renal function and stage of HIV infection on post-treatment changes in renal function among HIV-infected patients: a retrospective cohort study

OBJECTIVES

The use of combination antiretroviral therapy (cART) increases clinical uncertainty about changes in renal function. Specifically, little is known regarding the interaction of the effects of aging, baseline renal impairment, and stages of HIV infection on post-treatment changes in renal function.

METHODS

This analysis included 5533 HIV-infected patients on cART in 2004-2016. Progression to chronic kidney disease (CKD) was defined as either two consecutive estimated glomerular filtration rate (eGFR) measurements < 60 mL/min/1.73 m² for baseline eGFR ≥ 60 mL/min/1.73 m² (mild renal impairment or normal renal function) or a 25% decline for baseline eGFR < 60 mL/min/1.73 m² (moderate renal impairment).

RESULTS

During follow-up (median 4.8 years), 130 (2.3%) of the patients progressed to CKD. A total of 20.1% of patients with baseline normal renal function progressed to mild renal impairment, while 74.0% of patients with baseline mild or moderate renal impairment improved to normal renal function. In multivariable analysis, a significant positive baseline-eGFR-by-World Health Organization (WHO)-stage interaction effect on progression to CKD in all patients was identified, indicating a cross-over effect from a reduced risk to an increased risk. A significant negative baseline-age-by-WHO-stage interaction effect on progression to mild renal impairment in patients with baseline normal renal function was identified, with adjusted hazard ratios progressively lower at older ages. In addition, there were significant associations with older age, lower baseline eGFR, Dai ethnic minority, and anaemia for both outcomes, hyperglycaemia for CKD only, and higher CD4 count, tenofovir and ritonavir-boosted lopinavir use for mild renal impairment only.

CONCLUSIONS

Our data suggest a complex pattern of renal function dynamics in patients on cART, which requires precise management with systematic monitoring of the interaction of the effects of sociodemographic, nephrological and HIV-specific clinical characteristics.

[Long non-coding RNA HULC affects downstream-related targets to regulate migration and invasion of hepatoma cells]

Objective: To investigate the effects of long non-coding RNA HULC on downstream related targets regulating the migration and invasion of hepatoma cells and theirs mechanism of action. Methods: The expression of highly upregulated in liver cancer (HULC) in hepatocellular carcinoma, and adjacent normal liver tissues and different hepatocellular carcinoma cells were detected by qPCR. The correlation between clinicopathological data of HULC and liver cancer patients were analyzed. Dual-luciferase reporter gene detected the interaction between HULC and miR-186. CCK-8 assay was used to detect the effect of HULC on proliferation of hepatocellular carcinoma cells. The change in hepatocellular carcinoma cell invasions ability after HULC inhibition was detected by transwell invasion assay and migration ability after inhibition of HULC was assessed by scratch assay. Differences between groups were compared using one-way ANOVA. P < 0.05 was considered statistically significant. Results: Compared with adjacent normal liver tissue, the expression of HULC in hepatocellular carcinoma was significantly higher [(1.79 ± 0.25) vs. (0.23 ± 0.05), P < 0.05]. The expression level of HULC was highest in hepatocellular carcinoma HepG3 cells. HULC specifically banded to the 3'UTR of miR-186 and regulated the expressional activity of miR-186. After inhibiting the expression of HULC, the proliferation of hepatocellular carcinoma cells was 72 h (0.35 ± 0.09) vs. (0.82 ± 0.16), P < 0.05; 96 h (0.42 ± 0.08) vs.(1.28 ± 0.19), P < 0.05), and the ability of migration and invasion was relatively decreased in 24 h (11.2% ± 1.6%) vs. (23.5% ± 3.6%), P < 0.05; 48 h (18.6% ± 3.0%) vs. (38.6% ± 5.6%), P < 0.05; 72 h (43.6% ± 5.3% ) vs. (69.6% ± 7.6%), P < 0.05]. After inhibiting the expression of HULC, the tumor volume and body weight of tumor-bearing mice were significantly reduced [volume (2.89 ± 0.29) cm(3) vs. (0.89 ± 0.18) cm(3), P < 0.05, body weight (3.18 ± 0.41) g vs. (0.45 ± 0.09) g, P < 0.05]. Conclusion: HULC plays an important role in the occurrence and development of hepatocellular carcinoma and can influence the biological behavior of hepatoma cells by regulating the expression of downstream-related targets.

DNAH6 is a novel candidate gene associated with sperm head anomaly

Globozoospermia and acephalic spermatozoa are two rare sperm head anomalies associated with male infertility. Combination of the two phenotypes in the same patient is extremely rare, so the underlying pathogenesis of this disorder remains unclear. Here, we report a 35-year-old infertile male, who presented with 30% of sperm-lacked heads and 69% of sperm round-headed or small-headed with neck thickening in his ejaculate. Subsequent whole-exome sequencing (WES) analysis identified compound heterozygous variants within the DNAH6 gene. DNAH6 is a testis-specific-expressed protein that was localised to the neck region in the spermatozoa of normal control; however, immunofluorescent staining failed to detect DNAH6 protein in the patient's spermatozoa. Quantitative real-time PCR analysis also showed the complete absence of DNAH6 mRNA in the patient's spermatozoa. Moreover, two cycles of in vitro fertilisation (IVF)-assisted reproduction were carried out, but pregnancy was not achieved after embryo transfer. Therefore, rare sequence variants in DNAH6 might be susceptibility risks for human sperm head anomaly.