Applying deep learning for adverse pregnancy outcome detection with pre-pregnancy health data

Adverse pregnancy outcomes can bring enormous losses to both families and the society. Thus, pregnancy outcome prediction stays a crucial research topic as it may help reducing birth defect and improving the quality of population. However, recent advances in adverse pregnancy outcome detection are driven by data collected after mothers having been pregnant. In this situation, if a bad pregnancy outcome is diagnosed, the parents will suffer both physically and emotionally. In this paper, we develop a deep learning algorithm which is able to detect and classify adverse pregnancy outcomes before parents getting pregnant. We train a multi-layer neural network by using a dataset of 75542 couples’ multidimension pre-pregnancy health data. Our model outperforms some of algorithms in accuracy, recall and F1 score.

Download Full-text

Reproductive Concerns Among Female Leukemia and Lymphoma Survivors with the History of Adverse Pregnancy Outcomes after Cancer Treatment

Blood ◽

10.1182/blood-2018-99-110590 ◽

2018 ◽

Vol 132 (Supplement 1) ◽

pp. 5893-5893

Author(s):

Ksenya Shliakhtsitsava ◽

Emily Myers ◽

Irene Su

Keyword(s):

Pregnancy Outcome ◽

Pregnancy Outcomes ◽

Adverse Pregnancy Outcome ◽

Adverse Pregnancy Outcomes ◽

The Future ◽

Reproductive Concerns ◽

History Of ◽

Adverse Pregnancy ◽

After Cancer ◽

Lymphoma Survivors

Abstract Introduction Due to advances in cancer treatment the majority of young adults diagnosed with leukemia or lymphoma become long-term survivors. These individuals have been shown to have higher pregnancy risks as a result of cancer and treatment exposure. Adverse pregnancy outcomes including miscarriage and premature delivery may impact survivor's reproductive concerns after cancer. We hypothesized that reproductive-aged female leukemia and lymphoma survivors who experienced miscarriage or premature birth after cancer would have higher reproductive concerns as compared to female leukemia and lymphoma survivors who did not have an adverse pregnancy outcome after cancer. Methods This is a retrospective cohort study of young adult female leukemia and lymphoma survivors with at least one pregnancy after cancer, who are participants of the Reproductive Window study. Study participants were recruited between March 2015 and December 2017 from population-based cancer registries (California and Texas), physician and advocacy group referrals. Eligible women were age 18 to 40 at enrollment, age 15-35 at cancer diagnosis, and had at least one ovary. Enrolled participants answered a questionnaire on pregnancy outcomes and reproductive concerns using the Reproductive Concerns After Cancer (RCAC) Scale. The exposure was adverse pregnancy outcome (miscarriage or premature birth). The outcomes were RCAC subscales measuring concerns regarding becoming pregnant in the future and personal and offspring health. Subscale scores were dichotomized at 3, with >3 indicating moderate to severe reproductive concerns. Logistic regression models were used to test the association between an adverse pregnancy outcome (miscarriage or preterm birth) and RCAC subscales of interest, while adjusting for confounding Results 76 participants, mean age 34.3±3.9 years and mean years since cancer diagnosis 12.0± 5.8 years were included. The majority of participants were white (80%), completed college (72%) and were partnered (87%). Thirty eight percent of participants reported an adverse pregnancy outcome after cancer (18% miscarriage, 21% premature delivery). Thirty-two percent reported moderate to severe concerns about becoming pregnant in the future, 60% regarding offspring health, and 46% over personal health. History of miscarriage after cancer, but not preterm birth, was associated with higher concerns about becoming pregnant. In a multivariable model that adjusted for Hispanic ethnicity, current age, and live birth after cancer, participants who experienced a miscarriage after cancer were more likely to have moderate to high concerns about becoming pregnant in the future (adjusted OR 4.1, 95% CI 1.05-15.5, p=0.042) compared to participants with no history of miscarriage. Neither adverse pregnancy outcomes was associated with concerns about offspring or personal health. Conclusions In the cohort of young adult female leukemia and lymphoma survivors, experiencing a miscarriage after cancer was associated with higher concerns regarding becoming pregnant in the future. Additional research is needed to determine whether interventions such as preconception counseling with consideration of prior cancer treatments may help manage these concerns and improve pregnancy outcomes. Disclosures No relevant conflicts of interest to declare.

Download Full-text

Effects of sanitation practices on adverse pregnancy outcomes in India: a conducive finding from recent Indian demographic health survey

BMC Pregnancy and Childbirth ◽

10.1186/s12884-019-2528-8 ◽

2019 ◽

Vol 19 (1) ◽

Cited By ~ 1

Author(s):

Ratna Patel ◽

Ajay Gupta ◽

Shekhar Chauhan ◽

Dhananjay W. Bansod

Keyword(s):

Birth Weight ◽

Low Birth Weight ◽

Pregnancy Outcome ◽

Health Survey ◽

Pregnancy Outcomes ◽

Adverse Pregnancy Outcome ◽

Wealth Index ◽

Adverse Pregnancy Outcomes ◽

Adverse Outcomes ◽

Adverse Pregnancy

Abstract Background Several risk factors predisposing women and their live-borns to adverse outcomes during pregnancy have been documented. Little is known about sanitation being a factor contributing to adverse pregnancy outcomes in India. The role of sanitation in adverse pregnancy outcomes remains largely unexplored in the Indian context. This study is an attempt to bring the focus on sanitation as a factor in adverse pregnancy outcome. Along with the sanitation factors, few confounder variables have also been studied in order to understand the adverse pregnancy outcomes. Methods The study is based on the fourth round of National Family Health Survey (NFHS-IV) covering 26,972 married women in the age-group 15–49. The study variables include the mother’s age, Body Mass Index (BMI), education, anemia, and Antenatal care (ANC) visits during their last pregnancy. The household level variable includes place of residence, religion, caste, wealth index, access to toilet, type of toilet, availability of water within toilet premises, and facility of hand wash near the toilet. Children study variables include Low Birth Weight (LBW), the order of birth (Parity), and the death of the children of the women in the last 5 years. The target variable Adverse Pregnancy Outcome (APO) was constructed using children born with low birth weight or died during the last pregnancy. Results We calculated both adjusted as well as unadjusted odds ratios for a better understanding of the association between sanitation and adverse pregnancy outcomes. Findings from the study showed that women who did not have access to a toilet within the house had a higher risk of adverse pregnancy outcome. In the multivariable model, no association was observed for adverse pregnancy outcome among women who did not have access to toilet and women who used shared toilet. Teenage (15–19 years), uneducated, underweight and anemic mothers were more likely to face APO as compare to other mothers in similar characteristics group. Conclusions Our findings contribute to the decidedly less available literature on maternal sanitation behaviour and adverse pregnancy outcomes. Our results support that sanitation is a very significant aspect for women who are about to deliver a baby as there was an association between sanitation and adverse pregnancy outcome. Education on sanitation practices is the need of the hour as much as it needs to follow.

Download Full-text

Treatment and pregnancy outcomes of pregnant women exposed to second-line anti-tuberculosis drugs in South Africa

BMC Pregnancy and Childbirth ◽

10.1186/s12884-021-03956-6 ◽

2021 ◽

Vol 21 (1) ◽

Author(s):

Idah Mokhele ◽

Nelly Jinga ◽

Rebecca Berhanu ◽

Thandi Dlamini ◽

Lawrence Long ◽

...

Keyword(s):

South Africa ◽

Pregnant Women ◽

Pregnancy Outcome ◽

Pregnancy Outcomes ◽

Adverse Pregnancy Outcome ◽

Adverse Pregnancy Outcomes ◽

Close Monitoring ◽

Second Line ◽

Tb Treatment ◽

Adverse Pregnancy

Abstract Background Multi-drug resistant and rifampicin-resistant tuberculosis (MDR/RR-TB) in pregnant women is a cause for concern globally; few data have described the safety of second-line anti-TB medications during pregnancy. We aim to describe TB treatment and pregnancy outcomes among pregnant women receiving second-line anti-tuberculosis treatment for MDR/RR-TB in Johannesburg, South Africa. Methods We conducted a retrospective record review of pregnant women (≥ 18 years) who received treatment for MDR/RR-TB between 01/2010–08/2016 at three outpatient treatment sites in Johannesburg, South Africa. Demographic, treatment and pregnancy outcome data were collected from available medical records. Preterm birth (< 37 weeks), and miscarriage were categorized as adverse pregnancy outcomes. Results Out of 720 women of child-bearing age who received MDR/RR-TB treatment at the three study sites, 35 (4.4%) pregnancies were identified. Overall, 68.7% (24/35) were HIV infected, 83.3% (20/24) were on antiretroviral therapy (ART). Most women, 88.6% (31/35), were pregnant at the time of MDR/RR-TB diagnosis and four women became pregnant during treatment. Pregnancy outcomes were available for 20/35 (57.1%) women, which included 15 live births (11 occurred prior to 37 weeks), 1 neonatal death, 1 miscarriage and 3 pregnancy terminations. Overall, 13/20 (65.0%) women with known pregnancy outcomes had an adverse pregnancy outcome. Of the 28 women with known TB treatment outcomes 17 (60.7%) completed treatment successfully (4 were cured and 13 completed treatment), 3 (10.7%) died and 8 (28.6%) were lost-to-follow-up. Conclusions Pregnant women with MDR/RR-TB suffer from high rates of adverse pregnancy outcomes and about 60% achieve a successful TB treatment outcome. These vulnerable patients require close monitoring and coordinated obstetric, HIV and TB care.

Download Full-text

Vitamin D Concentration during Early Pregnancy and Adverse Outcomes among HIV-Negative Women in Dar-es-Salaam, Tanzania: A Case-Control Study

Nutrients ◽

10.3390/nu11122906 ◽

2019 ◽

Vol 11 (12) ◽

pp. 2906

Author(s):

Aneth V. Kalinjuma ◽

Anne Marie Darling ◽

Christopher R. Sudfeld ◽

Ferdinand Mugusi ◽

Julie Wright ◽

...

Keyword(s):

Vitamin D ◽

Pregnancy Outcome ◽

Early Pregnancy ◽

Pregnancy Outcomes ◽

Case Control Study ◽

Adverse Pregnancy Outcome ◽

Adverse Pregnancy Outcomes ◽

Adverse Pregnancy ◽

Control Study ◽

Hiv Negative

We examined the associations of plasma vitamin D concentration and adverse pregnancy outcomes among HIV-negative women in Dar-es-Salaam, Tanzania. We used an unmatched case-control study design, with 25-hydroxyvitamin D [25(OH)D] concentration assessed in the first trimester. Cases were individuals with adverse pregnancy outcomes, including stillbirth, premature birth, or small for gestational age births (SGA). Unconditional logistic regression and weighted logistic regression models were used to describe the associations of 25(OH)D concentration with the composite of adverse pregnancy outcome and individual adverse pregnancy outcomes, respectively. We included 310 cases and 321 controls. In controls, 5(2%) were vitamin D deficient (25(OH)D < 20 ng/mL), and 17(5%) had insufficient 25(OH)D concentration (20.0–29.9 ng/mL). Women with 25(OH)D < 20 ng/mL had 1.82 times the odds of occurrence of the composite adverse pregnancy outcome (OR = 1.82, 95% CI: 0.56–5.93; p = 0.32), however we noted a non-linear association between 25(OH)D concentration and adverse pregnancy outcome (p = 0.02). We found a 3-fold increased odds of stillbirth in women with low 25(OH)D concentration (OR = 3.11, 95% CI: 1.18–8.23, p = 0.02). Vitamin D concentration in early pregnancy may be an important factor in determining the course of pregnancy. Further research is needed to investigate whether the association of maternal 25(OH)D concentration in early pregnancy and stillbirth is causal.

Download Full-text

Treatment and Pregnancy Outcomes of Pregnant Women Exposed to Second-line Anti-tuberculosis Drugs in South Africa

10.21203/rs.3.rs-155326/v1 ◽

2021 ◽

Author(s):

Idah Mokhele ◽

Nelly Jinga ◽

Rebecca Berhanu ◽

Thandi Dlamini ◽

Lawrence Long ◽

...

Keyword(s):

South Africa ◽

Pregnant Women ◽

Pregnancy Outcome ◽

Pregnancy Outcomes ◽

Adverse Pregnancy Outcome ◽

Adverse Pregnancy Outcomes ◽

Close Monitoring ◽

Second Line ◽

Tb Treatment ◽

Adverse Pregnancy

Abstract Background:Multi-drug resistant and rifampicin-resistant tuberculosis (MDR/RR-TB) in pregnant women is a cause for concern globally; few data have described the safety of second-line anti-TB medications during pregnancy. We characterize maternal, and pregnancy outcomes for pregnant women receiving second-line anti-tuberculosis treatment for MDR/RR-TB.Methods: We conducted a retrospective record review of pregnant women (≥18 years) who received treatment for MDR/RR-TB between 01/2010–08/2016 at three outpatient treatment sites in Johannesburg, South Africa. Demographic, treatment and pregnancy outcome data were collected from available medical records. Preterm birth (<37 weeks), and miscarriage were categorized as adverse pregnancy outcomes.Results: Out of 720 women of child-bearing age who received MDR/RR-TB treatment at the three study sites, 35 (4.4%) pregnancies were identified. Overall, 68.7% (24/35) were HIV infected, 83.3% were on ART. Most women (88.6%) were pregnant at the time of MDR/RR-TB diagnosis and four women became pregnant during treatment. Pregnancy outcomes were available for 20/35 (57.1%) women, which included 15 live births (11 occurred prior to 37 weeks), 1 neonatal death, 1 miscarriage and 3 pregnancy terminations. Overall, 13/17 (76.5%) had an adverse pregnancy outcome. Of the 28 women with known TB treatment outcomes 17 (60.7%) completed treatment successfully (4 were cured and 13 completed treatment), 3 (10.7%) died and 8 (28.6%) were lost-to-follow-up.Conclusions:Pregnant women with MDR/RR-TB suffer from high rates of adverse pregnancy outcomes and about 60% achieve a successful TB treatment outcome. These vulnerable patients require close monitoring and coordinated obstetric, HIV and TB care.

Download Full-text

Neonatal outcome in pregnant patients with antiphospholipid syndrome

Journal of Perinatal Medicine ◽

10.1515/jpm-2014-0118 ◽

2015 ◽

Vol 43 (6) ◽

Cited By ~ 13

Author(s):

Katarina Jeremic ◽

Aleksandar Stefanovic ◽

Jelena Dotlic ◽

Jelena Stojnic ◽

Sasa Kadija ◽

...

Keyword(s):

Risk Factors ◽

Respiratory Distress Syndrome ◽

Pregnancy Outcome ◽

Pregnancy Outcomes ◽

Distress Syndrome ◽

Adverse Pregnancy Outcome ◽

Perinatal Asphyxia ◽

Premature Delivery ◽

Control Group ◽

Adverse Pregnancy

AbstractThe study aim was to evaluate pregnancy outcomes in patients with antiphospholipid syndrome (APS) and to determine which clinical parameters present risk factors for adverse pregnancy outcomes in these patients.The study included 55 patients with APS treated at the Clinic for Ob/Gyn, Clinical Center of Serbia, from 2006 to 2012. The control group consisted of 55 healthy pregnant women. Data regarding previous pregnancies and conception method were registered. Immunological and laboratory tests were performed. Pregnancy outcomes, including miscarriage, intrauterine fetal death, hypertensive disorders, diabetes mellitus, phlebothrombosis, fetal growth restriction, premature delivery, delivery method, perinatal asphyxia, respiratory distress syndrome, intraventricular hemorrhage, and necrotizing enterocolitis were followed.The premature delivery rate in APS patients was 31.8%, and pregnancy loss was 18.2%. Significantly more patients with APS had thrombocytopenia, pregnancy losses, intrauterine growth restriction, and perinatal asphyxia compared with the control group. More miscarriages, preterm delivery, lower birth weight, preeclampsia, and IgM anticardiolipin antibody levels significantly correlated with adverse pregnancy outcomes. Although rare, respiratory distress syndrome can also worsen neonatal health status. According to ROC analysis, previous miscarriages correctly explained 66.3% of adverse pregnancy outcome cases. We generated four equations of adverse pregnancy outcome risk factors.The most important prognostic factor for pregnancy outcome in APS patients is the number of previous miscarriages. Using appropriate current therapeutic protocol can enable live birth of a healthy newborn in most cases.

Download Full-text

Potential value of the PixelShine deep learning algorithm for increasing quality of 70 kVp+ASiR-V reconstruction pelvic arterial phase CT images

Japanese Journal of Radiology ◽

10.1007/s11604-018-0798-0 ◽

2018 ◽

Vol 37 (2) ◽

pp. 186-190 ◽

Cited By ~ 6

Author(s):

Shi-feng Tian ◽

Ai-lian Liu ◽

Jing-hong Liu ◽

Yi-jun Liu ◽

Ju-dong Pan

Keyword(s):

Deep Learning ◽

Learning Algorithm ◽

Ct Images ◽

Arterial Phase ◽

Potential Value ◽

Deep Learning Algorithm

Download Full-text

Abstract P176: Associations of Gestational Lipids and Apolipoproteins With Pregnancy Outcomes: The Hyperglycemia and Adverse Pregnancy Outcome Study

Circulation ◽

10.1161/circ.141.suppl_1.p176 ◽

2020 ◽

Vol 141 (Suppl_1) ◽

Author(s):

Amanda M Perak ◽

Alan Kuang ◽

Nicola Lancki ◽

Darwin R Labarthe ◽

Svati H Shah ◽

...

Keyword(s):

Blood Pressure ◽

Pregnancy Outcome ◽

Gestational Age ◽

Pregnancy Outcomes ◽

Adverse Pregnancy Outcome ◽

Late Pregnancy ◽

Large For Gestational Age ◽

Obstetric Outcomes ◽

Outcome Study ◽

Adverse Pregnancy

Introduction: Gestational hyperlipidemia has traditionally been considered physiologic and benign, but the significance of inter-individual variation in lipid levels for maternal-fetal health are poorly understood. We examined associations of gestational lipids and apolipoproteins with adverse obstetric and neonatal outcomes. Methods: Data from the Hyperglycemia and Adverse Pregnancy Outcome Study were analyzed, including 1,813 mother-child dyads from 9 field centers in 6 countries: US (25%), Barbados (24%), UK (20%), China (16%), Thailand (8%), and Canada (7%). Fasting lipids and apolipoproteins were directly measured at a mean of 28 (range 23-34) weeks’ gestation. Cord blood was collected at delivery, neonatal anthropometrics were measured within 72 hours, and medical records were abstracted for obstetric outcomes. Logistic regression was utilized to test associations of lipids and apolipoproteins (per +1 SD; log-transformed if skewed) with pregnancy outcomes, adjusted for center, demographics, and maternal covariates such as BMI, blood pressure, and glycemia. Results: See Table for lipid and apolipoprotein levels in pregnant mothers. In fully adjusted models ( Table ), 1 SD higher log-triglycerides (i.e., ~2.7-fold higher triglyceride level) in late pregnancy was significantly associated with higher odds for preeclampsia (OR 1.53 [95% CI, 1.15-2.05]), large for gestational age infant (1.42 [1.21-1.67]), and infant insulin sensitivity <10 th percentile (1.25 [1.03-1.50]), but not with unplanned primary cesarean section or infant sum of skinfolds >90 th percentile. There were no significant associations of maternal HDL-C, LDL-C, or log-ApoB/A1 ratio with any outcome. Conclusion: Triglyceride levels in the latter half of pregnancy were uniquely associated with both maternal risks (preeclampsia) and neonatal risks (large for gestational age and insulin resistance), even after adjustment for maternal BMI, blood pressure, and glycemia.

Download Full-text