ABSTRACT
INTRODUCTION: Early prediction of preeclampsia (PE) is of universal importance in controlling the disease process. Our study aimed to assess the feasibility of using retinal fundus images to predict preeclampsia via deep learning in singleton pregnancies. METHODS: This prospective cohort study was conducted at Shanghai First Maternity and Infant Hospital, Tongji University School of Medicine. Eligible participants included singleton pregnancies who presented for prenatal visits before 14âweeks of gestation from September 1, 2020, to February 1, 2022. Retinal fundus images were obtained using a nonmydriatic digital retinal camera during their initial prenatal visit upon admission before 20âweeks of gestation. In addition, we generated fundus scores, which indicated the predictive value of hypertension, using a hypertension detection model. To evaluate the predictive value of the retinal fundus image-based deep learning algorithm for preeclampsia, we conducted stratified analyses and measured the area under the curve (AUC), sensitivity, and specificity. We then conducted sensitivity analyses for validation. RESULTS: Our study analyzed a total of 1138 women, 92 pregnancies developed into hypertension disorders of pregnancy (HDP), including 26 cases of gestational hypertension and 66 cases of preeclampsia. The adjusted odds ratio (aOR) of the fundus scores was 2.582 (95% CI, 1.883-3.616; P â<â0.001). Otherwise, in the categories of prepregnancy BMI less than 28.0 and at least 28.0, the aORs were 3.073 (95%CI, 2.265-4.244; P â<â0.001) and 5.866 (95% CI, 3.292-11.531; P â<â0.001). In the categories of maternal age less than 35.0 and at least 35.0, the aORs were 2.845 (95% CI, 1.854-4.463; P â<â0.001) and 2.884 (95% CI, 1.794-4.942; P â<â0.001). The AUC of the fundus score combined with risk factors was 0.883 (sensitivity, 0.722; specificity, 0.934; 95% CI, 0.834-0.932) for predicting preeclampsia. CONCLUSION: Our study demonstrates that the use of deep learning algorithm-based retinal fundus images offers promising predictive value for the early detection of preeclampsia.
Subject(s)
Deep Learning , Hypertension, Pregnancy-Induced , Pre-Eclampsia , Female , Pregnancy , Humans , Pre-Eclampsia/diagnostic imaging , Prospective Studies , China , Hypertension, Pregnancy-Induced/diagnosisABSTRACT
OBJECTIVE: To investigate reliable biomarkers for predicting histological chorioamnionitis (HCA) in women with preterm prelabour rupture of membranes (PPROM). DESIGN: A retrospective study. SETTING: A maternity care hospital in Shanghai. POPULATION: Women with PPROM before 34+0/7 weeks of gestation. METHODS: Mean values of biomarkers were compared by two-way analysis of variance (ANOVA). Log-binomial regression models were used to assess the association between biomarkers and risk of HCA. A stepwise logistic regression model was used to develop a multi-biomarker prediction model and identify the independent predictors. The area under the receiver operating characteristic curve (AUC) was used to assess prediction performance. MAIN OUTCOME MEASURES: The ability of the individual biomarker and the combination of multiple biomarkers to predict HCA. RESULTS: In 157 mothers with PPROM, 98 (62.42%) women had HCA and 59 (37.58%) women did not have HCA. No significant differences were observed between the two groups in white blood cell, neutrophil or lymphocyte counts, whereas both high-sensitivity C-reactive protein (hsCRP) and procalcitonin (PCT) were significantly higher in the HCA group. HsCRP and PCT were found to be independently associated with the risk of HCA, and PCT had a larger AUC value than hsCRP (p < 0.05). The optimal multi-biomarker prediction model for HCA (AUC = 93.61%) included hsCRP at 72 hours and PCT at 48 and 72 hours, and PCT had a stronger prediction capacity than hsCRP. CONCLUSIONS: PCT could be a reliable biomarker for the early prediction of HCA in women with PPROM within 72 hours of dexamethasone treatment.
Subject(s)
Chorioamnionitis , Fetal Membranes, Premature Rupture , Maternal Health Services , Infant, Newborn , Female , Pregnancy , Humans , Male , Chorioamnionitis/diagnosis , Retrospective Studies , C-Reactive Protein/analysis , China/epidemiology , Biomarkers , DexamethasoneABSTRACT
OBJECTIVE: We aimed to evaluate the prognostic value of C-C motif chemokine receptor type 5 (CCR5) expression level for patients with ovarian cancer and to establish a radiomics model that can predict CCR5 expression level using The Cancer Imaging Archive (TCIA) and The Cancer Genome Atlas (TCGA) database. METHODS: A total of 343 cases of ovarian cancer from the TCGA were used for the gene-based prognostic analysis. Fifty seven cases had preoperative computed tomography (CT) images stored in TCIA with genomic data in TCGA were used for radiomics feature extraction and model construction. 89 cases with both TCGA and TCIA clinical data were used for radiomics model evaluation. After feature extraction, a radiomics signature was constructed using the least absolute shrinkage and selection operator (LASSO) regression analysis. A prognostic scoring system incorporating radiomics signature based on CCR5 expression level and clinicopathologic risk factors was proposed for survival prediction. RESULTS: CCR5 was identified as a differentially expressed prognosis-related gene in tumor and normal sample, which were involved in the regulation of immune response and tumor invasion and metastasis. Four optimal radiomics features were selected to predict overall survival. The performance of the radiomics model for predicting the CCR5 expression level with 10-fold cross- validation achieved Area Under Curve (AUCs) of 0.770 and of 0.726, respectively, in the training and validation sets. A predictive nomogram was generated based on the total risk score of each patient, the AUCs of the time-dependent receiver operating characteristic (ROC) curve of the model was 0.8, 0.673 and 0.792 for 1-year, 3-year and 5-year, respectively. Along with clinical features, important imaging biomarkers could improve the overall survival accuracy of the prediction model. CONCLUSION: The expression levels of CCR5 can affect the prognosis of patients with ovarian cancer. CT-based radiomics could serve as a new tool for prognosis prediction.
Subject(s)
Ovarian Neoplasms , Tomography, X-Ray Computed , Humans , Female , Retrospective Studies , Prognosis , Tomography, X-Ray Computed/methods , Machine Learning , Ovarian Neoplasms/diagnostic imaging , Ovarian Neoplasms/genetics , Receptors, CCR5/geneticsABSTRACT
BACKGROUND: Frozen-thawed embryo transfer (FET) is thought to be associated with obstetric and neonatal complications after in vitro fertilization/intracytoplasmic single sperm injection (IVF/ICSI) treatment. The study aimed to determine whether the endometrial preparation protocol is an influencing factor for these complications. METHODS: We conducted a retrospective cohort study of 3,458 women who had singleton deliveries after IVF/ICSI-FET treatment at the Centre for Reproductive Medicine of Shanghai First Maternity and Infant Hospital between July 2016 and April 2021. The women were divided into three groups according to the endometrial preparation protocols: 2,029 women with programmed cycles, 959 with natural cycles, and 470 with minimal ovarian stimulation cycles. The primary outcomes were the incidence rates of obstetric and neonatal complications, namely, hypertensive disorders of pregnancy (HDP), gestational diabetes mellitus (GDM), intrahepatic cholestasis of pregnancy (ICP), placenta previa, preterm rupture of membranes (PROM), preterm delivery, postpartum haemorrhage, large for gestational age (LGA), small for gestational age (SGA), and macrosomia. RESULTS: After adjustments for confounding variables by multivariate logistic regression analysis, the results showed that programmed cycles had an increased risk of HDP (aOR = 1.743; 95% CI, 1.110-2.735; P = 0.016) and LGA (aOR = 1.269; 95% CI, 1.011-1.592; P = 0.040) compared with natural cycles. Moreover, programmed cycles also increased the risk of LGA (aOR = 1.459; 95% CI, 1.083-1.965; P = 0.013) but reduced the risk of SGA (aOR = 0.529; 95% CI, 0.348-0.805; P = 0.003) compared with minimal ovarian stimulation cycles. There were no significant differences between natural cycles and minimal ovarian stimulation cycles. CONCLUSIONS: During IVF/ICSI-FET treatment, the risk of HDP and LGA was increased in women with programmed cycles. Therefore, for patients with thin endometrium, irregular menstruation or no spontaneous ovulation, minimal ovarian stimulation cycles may be a relatively safer option than programmed cycles.
Subject(s)
Embryo Transfer , Semen , China/epidemiology , Cryopreservation , Embryo Transfer/adverse effects , Endometrium , Female , Fertilization in Vitro/adverse effects , Humans , Infant, Newborn , Male , Pregnancy , Pregnancy Outcome/epidemiology , Retrospective StudiesABSTRACT
Background: Venous thromboembolism (VTE) remains an important cause of maternal deaths. Little is known about the associations of specific periods of gestational weight gain (GWG) with the category of VTE, pulmonary embolism (PE), or deep venous thrombosis (DVT) with or without PE. Methods: In a retrospective case-control study conducted in Shanghai First Maternity and Infant Hospital from January 1, 2017 to September 30, 2021, cases of VTE within pregnancy or the first 6 postnatal weeks were identified. Controls without VTE were randomly selected from women giving birth on the same day as the cases, with 10 controls matched to each case. Total GWG and rates of early, mid, and late GWG values were standardized into z-scores, stratified by pre-pregnant body mass index (BMI). The adjusted odds ratios (aORs) and 95% confidence intervals (CIs) were estimated through multivariate logistic regression models. Results: There were 196 cases (14.4 per 10,000) of VTE within pregnancy or the first 6 postnatal weeks were identified. Higher total weight gain was associated with increased risks of PE (aOR, 13.22; 95% CI, 2.03-85.99) and VTE (OR, 10.49; 95% CI, 1.82-60.45) among women with underweight. In addition, higher total weight gain was associated with increased risk of PE (aOR, 2.06; 95% CI, 1.14-3.72) among women with healthy weight. Similarly, rate of higher early weight gain was associated with significantly increased risk for PE (aOR, 2.15; 95% CI, 1.05-4.42) among women with healthy BMI. The lower rate of late weight gain was associated with increased risks of PE (aOR, 7.30; 95% CI, 1.14-46.55) and VTE (OR, 7.54; 95% CI, 1.20-47.57) among women with underweight. No significant associations between maternal rate of mid GWG and increased risk for any category of VTE, PE, or DVT with or without PE were present, regardless of maternal pre-pregnant BMI. Conclusion: The GWG associations with the category of VTE, PE, or DVT with or without PE differ at different periods of pregnancy. In order to effectively improve maternal and child outcomes, intensive weight management that continues through pregnancy may be indispensable.