Artificial Intelligence to Determine Fetal Sex.

OBJECTIVE: This proof-of-concept study assessed how confidently an artificial intelligence (AI) model can determine the sex of a fetus from an ultrasound image. STUDY DESIGN: Analysis was performed using 19,212 ultrasound image slices from a high-volume fetal sex determination practice. This dataset was split into a training set (11,769) and test set (7,443). A computer vision model was trained using a transfer learning approach with EfficientNetB4 architecture as base. The performance of the computer vision model was evaluated on the hold out test set. Accuracy, Cohen's Kappa and Multiclass Receiver Operating Characteristic area under the curve (AUC) were used to evaluate the performance of the model. RESULTS: The AI model achieved an Accuracy of 88.27% on the holdout test set and a Cohen's Kappa score 0.843. The ROC AUC score for Male was calculated to be 0.896, for Female a score of 0.897, for Unable to Assess a score of 0.916, and for Text Added a score of 0.981 was achieved. CONCLUSION: This novel AI model proved to have a high rate of fetal sex capture that could be of significant use in areas where ultrasound expertise is not readily available. KEY POINTS: · This is the first proof-of-concept AI model to determine fetal sex.. · This study adds to the growing research in ultrasound AI.. · Our findings demonstrate AI integration into obstetric care..

Stillbirth in women with diabetes: a retrospective analysis of fetal autopsy reports.

INTRODUCTION: Diabetes in pregnancy is associated with an increased rate of stillbirth. There are a wide variety of factors that have been implicated including placental insufficiency, hypoxia, hyperinsulinemia and impaired cardiac function. Furthermore, there is evidence that diabetic pregnancies have an increased rate of fetal cardiomyopathy as compared to non-diabetic pregnancies. Prior studies have indicated that this association can also be an etiology for diabetic stillbirth. The purpose of this study was to determine if diabetic pregnancies have an increased risk of fetal cardiomyopathy identified on fetal autopsy as compared to non-diabetic women with a stillbirth in a cohort of pregnancies that had evaluation with a fetal autopsy. MATERIALS AND METHODS: Retrospective cohort study of women with a stillbirth who consented to fetal autopsy at an academic medical center from 2011 to 2017. Stillbirth was defined as an intrauterine fetal demise at ≥20 weeks' gestation. Women with diabetes defined as pre-gestational diabetes type 1, pre-gestational diabetes type 2, and gestational diabetes were compared to women without diabetes. Primary outcome was fetal cardiomyopathy. Other etiologies for stillbirth were also evaluated and classified according to the Stillbirth Collaborative Research Network (SCRN) initial causes of fetal death. Fisher exact test, χ2 test, and Mann Whitney U tests were performed as appropriate, with p < .05 considered significant. Generalized linear models were performed for fetal organ weights controlling for gestational age of delivery, maternal chronic hypertension, delivery body mass index, and birthweight. RESULTS: A total of 78 women elected to have fetal autopsy examinations during the study period. Of these, 75 had complete information available for review. A total of 60 women did not have diabetes and 15 women had diabetes. Of pregnancies complicated by diabetes, 11 had insulin dependent diabetes and 4 had non-insulin dependent diabetes. Fetal cardiomyopathy was diagnosed on autopsy for 7 (46.7%) of pregnancies with diabetes and 2 (3.3% of pregnancies without diabetes (RR 14.00 [95% CI 3.23-60.65], p < .001). These associations were still significant even when analyzing only those pregnancies without fetal congenital heart disease (7 [46.7%] diabetic pregnancies with cardiomyopathy versus 1 [2.0%] nondiabetic pregnancy with cardiomyopathy, RR 23.80 [95% CI 3.17-178.46], p < .001). There was no difference between diabetic and non-diabetic pregnancies in regards to other causes for stillbirth. Stillbirths in pregnancies with diabetes also had larger fetal heart, liver, and adrenal weights on fetal autopsy. CONCLUSION: Women with diabetes have 14 times the risk of fetal cardiomyopathy identified at fetal autopsy as compared to women without diabetes. As the prediction and prevention of diabetic stillbirth is limited, information on potential causes of stillbirth may help future research identify those pregnancies at the greatest risk for adverse outcome.

Preterm prelabor rupture of membranes: evaluating latency and neonatal morbidity for pregnancies with expectant management ≥34 weeks.

OBJECTIVE: To evaluate the association between latency and neonatal morbidity for pregnancies with expectant management of PPROM ≥34 w. MATERIALS AND METHODS: A retrospective cohort of singletons with PPROM from 2011 to 2016. Exposure was defined as latency (period from diagnosis of PPROM to delivery) and was analyzed as a count variable (i.e. number of days) and binary variable (≥7 days and <7 days; ≥21 days and <21 days). Primary outcome was composite neonatal morbidity defined as need for respiratory support, culture positive neonatal sepsis, and/or neonatal antibiotics >72 h. Fisher's exact test, chi-square test, Mann-Whitney U and binary logistic regression tests were performed with p<.05 considered significant. RESULTS: Of 813 pregnancies, 104 met inclusion criteria: 73 (70.2%) pregnancies with PPROM diagnosed at <34 weeks and 31 (29.8%) pregnancies with PPROM diagnosed ≥34 weeks. A total of 58 (55.8%) pregnancies had a latency of ≥7 days and 46 (44.2%) had a latency <7 days. There was no difference in composite neonatal morbidity for latency ≥7 d versus <7 d (aOR 0.92; 95% CI 0.30-2.82) or latency as a count variable (aOR 0.70; 95% CI 0.23-2.13). However, a latency ≥21 d was associated with increased composite neonatal morbidity (aOR 10.24, 95% CI 1.42-73.99). CONCLUSION: In pregnancies with PPROM expectantly managed ≥34 w, a latency of ≥7 d is not associated with significant differences in neonatal morbidity. However, different latency thresholds may be more clinically relevant for late preterm pregnancies. The increase in composite neonatal morbidity associated with a latency >21 days should be an area of future investigation and may suggest there is a population of pregnancies with PPROM which may not benefit from expectant management past 34 weeks.

Impact of Maternal Obesity on Perinatal Outcomes in Preterm Prelabor Rupture of Membranes ≥34 Weeks.

OBJECTIVE: This study aimed to compare pregnancy outcomes in obese and nonobese women with preterm prelabor rupture of membranes (PPROM) ≥34 weeks. STUDY DESIGN: The present study is a secondary analysis of a multicenter retrospective cohort of singletons with PPROM from 2011 to 2017. Women with a delivery body mass index (BMI) ≥30 kg/m2 (obese) were compared with women with a BMI < 30 kg/m2 (nonobese). Pregnancies were stratified based on delivery policies of expectant management until 35 weeks versus immediate delivery ≥34 weeks. The primary outcome was a composite neonatal outcome (neonatal sepsis, antibiotic administration for duration >72 hours after delivery or respiratory support). Univariate analysis and general estimating equations models including maternal age, delivery timing, mode of delivery, hospital, and gestational age were used with p < 0.05 level of significance. RESULTS: Among 259 pregnancies, 47% were obese. Pregnant women with obesity had increased composite neonatal outcome versus nonobese pregnancies (adjusted odds ratio [aOR] = 1.48 [95% confidence interval (CI): 1.01-2.17]). Obesity was also associated with increased neonatal antibiotic administration for a duration >72 hours after delivery, respiratory support, ventilation, oxygen supplementation, and surfactant administration. When stratified by delivery policies there was no significant difference in perinatal outcomes based on obesity. CONCLUSION: Obese women with PPROM ≥34 weeks have an increased odds of adverse neonatal respiratory and infectious outcomes compared with nonobese women.

Preterm Prelabor Rupture of Membranes: Outcomes with Expectant Management until 34 versus 35 Weeks.

OBJECTIVE: To evaluate outcomes with expectant management of preterm prelabor rupture of membranes (PROM) until 35 weeks versus immediate delivery at ≥34 weeks. STUDY DESIGN: This was a multicenter retrospective cohort study of singletons with preterm PROM at >20 weeks from 2011 through 2017. Groups were defined as expectant management until 35 weeks versus immediate delivery at ≥34 weeks. Primary outcome was composite neonatal morbidity: need for respiratory support, culture positive neonatal sepsis, or antibiotic administration for >72 hours. Univariate and general estimating equation models were used with p < 0.05 considered significant. RESULTS: A total of 280 mother-infant dyads were included. There was no difference in composite neonatal outcome in pregnancies managed with expectant management compared with immediate delivery (43.4 vs. 37.5%; p = 0.32). Those with expectant management had shorter length of neonatal intensive care unit (NICU) admission but higher rates of neonatal antibiotics for > 72 hours, endometritis, and histological chorioamnionitis. There were no cases of fetal demise, neonatal death, or maternal sepsis, and only three cases of neonatal sepsis. CONCLUSION: There is no difference in composite neonatal morbidity in pregnancies with preterm PROM managed with expectant management until 35 weeks as compared with immediate delivery at 34 weeks. Expectant management is associated with a decreased length of NICU admission but increased short-term infectious morbidity.

Angiogenic and Antiangiogenic Markers for Prediction and Risk Classification of Preeclampsia.

Preeclampsia is a pregnancy-specific hypertensive disorder with multisystem involvement and is a significant cause of obstetric morbidity and mortality worldwide. A major issue in the treatment of preeclampsia stems from its still significant rates of misclassification and misdiagnosis. Angiogenic factors have been speculated as a possible diagnostic modality due to a perceived imbalance in angiogenesis in preeclampsia. Factors currently studied include soluble fms-like protein kinase 1 and placental growth factor. Because of significant mortality associated with preeclampsia it is felt that both early and accurate diagnosis of preeclampsia is imperative if this disease process is to be treated.

Chronic granulomatous disease carrier with recurrent poor obstetric outcome.

BACKGROUND: Chronic granulomatous disease is a primary immunodeficiency disorder characterized by severe recurrent bacterial and fungal infections. Female carriers of the X-linked form of the disorder usually are unaffected and rarely have serious infections. CASE: A 22-year-old pregnant patient known to be a carrier of the X-linked form of chronic granulomatous disease had a history of chorioamnionitis during her two previous pregnancies. During her third pregnancy, she presented again with the same diagnosis, which resulted in delivery at 25 weeks of gestation. CONCLUSION: Carriers of chronic granulomatous disease should be monitored closely during pregnancy, as if they have the disease. To decrease the risk of infectious morbidity and mortality, obstetricians should have a low threshold for starting prophylactic antibiotics early during pregnancy, even if the patient is asymptomatic.