Detection of non-cardiac fetal abnormalities by ultrasound at 11-14 weeks: systematic review and meta-analysis.

OBJECTIVES: To assess diagnostic accuracy of 2D ultrasound at 11-14 weeks gestation as a screening test for individual fetal anomalies and identify screening factors impacting detection. METHODS: Systematic review and meta-analysis, developed and registered with PROSPERO (CRD42018111781). MEDLINE, EMBASE, Web of Science Core Collection and The Cochrane Library) were searched for studies evaluating the diagnostic accuracy of screening for 16 pre-specified, non-cardiac, congenital anomalies considered to be of interest to the early anomaly scan. We included prospective and retrospective studies from any healthcare setting and low risk, mixed risk and unselected populations. The reference standard was the detection of an anomaly on postnatal or post-mortem examination. Data were extracted to populate 2 x 2 tables and meta-analysis (random-effects model) undertaken to determine the diagnostic accuracy of screening for the pre-specified anomalies (individually and as a composite). Secondary analyses were performed to determine the impact of (1) imaging protocol (2) ultrasound modality (3) publication year and (4) index of sonographer suspicion at time of scan. Post-hoc secondary analysis was conducted to assess performance for studies from 2010. Risk of bias and quality assessment was undertaken for included studies using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2). RESULTS: From 5684 citations, 202 papers were identified as eligible and reviewed, resulting in the inclusion of 526,322 fetuses (52 studies) of which 2,399 were affected by one or more of the 16 anomalies. Individual anomalies were not equally amenable to detection on first trimester ultrasound ranging from high (>80%) detection rates for severe conditions including acrania (98%), gastroschisis (96%) and exomphalos (95%) and holoprosencephaly (88%); they were lower for open spina bifida (69%), lower urinary tract obstruction (66%) lethal skeletal dysplasias (57%) and limb reduction defects (50%) and below 50% for facial clefts (43%), polydactyly (40%) and congenital diaphragmatic hernia (38). Conditions with low (<30%) detection rates included bilateral renal agenesis (25%), closed spina bifida (21%), isolated cleft lip only (14%) and talipes (11%). Specificity was >99% for all anomalies. Secondary analysis showed improvement of detection with publication year, and that the use of imaging protocols had a statistically significant impact on screening performance (p<0.0001). CONCLUSIONS: Accurate detection of congenital anomalies using first trimester ultrasound is feasible. In this study we have determined screening characteristics for individual anomalies and have shown that detection rates and false positive rates are dependent on the type of anomaly. The use of a standardised protocol allows diagnostic performance to be maximised, and this particularly enhances screening performance for the detection of spina bifida, facial clefts and limb reduction defects. Highlighting the types of anomalies amenable to diagnosis and determining favourable screening test factors can support the development of first-trimester anomaly screening programs. This article is protected by copyright. All rights reserved.

Standardization and quality control of Doppler and fetal biometric ultrasound measurements in low-income setting.

OBJECTIVE: The aim of this study was to determine the quality of fetal biometry and pulsed-wave Doppler ultrasound measurements in a prospective cohort study in Uganda. METHODS: This was an ancillary study of the Ending Preventable Stillbirths by Improving Diagnosis of Babies at Risk (EPID) project, in which women enroled in early pregnancy underwent Doppler and fetal biometric assessment at 32-40 weeks of gestation. Sonographers undertook 6 weeks of training followed by onsite refresher training and audit exercises. A total of 125 images for each of the umbilical artery (UA), fetal middle cerebral artery (MCA), left and right uterine arteries (UtA), head circumference (HC), abdominal circumference (AC) and femur length (FL) were selected randomly from the EPID study database and evaluated independently by two experts in a blinded fashion using objective scoring criteria. Inter-rater agreement was assessed using modified Fleiss' kappa for nominal variables and systematic errors were explored using quantile-quantile (Q-Q) plots. RESULTS: For Doppler measurements, 96.8% of the UA images, 84.8% of the MCA images and 93.6% of the right UtA images were classified as of acceptable quality by both reviewers. For fetal biometry, 96.0% of the HC images, 96.0% of the AC images and 88.0% of the FL images were considered acceptable by both reviewers. The kappa values for inter-rater reliability of quality assessment were 0.94 (95% CI, 0.87-0.99) for the UA, 0.71 (95% CI, 0.58-0.82) for the MCA, 0.87 (95% CI, 0.78-0.95) for the right UtA, 0.94 (95% CI, 0.87-0.98) for the HC, 0.93 (95% CI, 0.87-0.98) for the AC and 0.78 (95% CI, 0.66-0.88) for the FL measurements. The Q-Q plots indicated no influence of systematic bias in the measurements. CONCLUSIONS: Training local healthcare providers to perform Doppler ultrasound, and implementing quality control systems and audits using objective scoring tools in clinical and research settings, is feasible in low- and middle-income countries. Although we did not assess the impact of in-service retraining offered to practitioners deviating from prescribed standards, such interventions should enhance the quality of ultrasound measurements and should be investigated in future studies. © 2022 The Authors. Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology.

Clinical workflow of sonographers performing fetal anomaly ultrasound scans: deep-learning-based analysis.

OBJECTIVE: Despite decades of obstetric scanning, the field of sonographer workflow remains largely unexplored. In the second trimester, sonographers use scan guidelines to guide their acquisition of standard planes and structures; however, the scan-acquisition order is not prescribed. Using deep-learning-based video analysis, the aim of this study was to develop a deeper understanding of the clinical workflow undertaken by sonographers during second-trimester anomaly scans. METHODS: We collected prospectively full-length video recordings of routine second-trimester anomaly scans. Important scan events in the videos were identified by detecting automatically image freeze and image/clip save. The video immediately preceding and following the important event was extracted and labeled as one of 11 commonly acquired anatomical structures. We developed and used a purposely trained and tested deep-learning annotation model to label automatically the large number of scan events. Thus, anomaly scans were partitioned as a sequence of anatomical planes or fetal structures obtained over time. RESULTS: A total of 496 anomaly scans performed by 14 sonographers were available for analysis. UK guidelines specify that an image or videoclip of five different anatomical regions must be stored and these were detected in the majority of scans: head/brain was detected in 97.2% of scans, coronal face view (nose/lips) in 86.1%, abdomen in 93.1%, spine in 95.0% and femur in 92.3%. Analyzing the clinical workflow, we observed that sonographers were most likely to begin their scan by capturing the head/brain (in 24.4% of scans), spine (in 23.2%) or thorax/heart (in 22.8%). The most commonly identified two-structure transitions were: placenta/amniotic fluid to maternal anatomy, occurring in 44.5% of scans; head/brain to coronal face (nose/lips) in 42.7%; abdomen to thorax/heart in 26.1%; and three-dimensional/four-dimensional face to sagittal face (profile) in 23.7%. Transitions between three or more consecutive structures in sequence were uncommon (up to 13% of scans). None of the captured anomaly scans shared an entirely identical sequence. CONCLUSIONS: We present a novel evaluation of the anomaly scan acquisition process using a deep-learning-based analysis of ultrasound video. We note wide variation in the number and sequence of structures obtained during routine second-trimester anomaly scans. Overall, each anomaly scan was found to be unique in its scanning sequence, suggesting that sonographers take advantage of the fetal position and acquire the standard planes according to their visibility rather than following a strict acquisition order. © 2022 The Authors. Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology.

Perinatal and long-term outcome of fetal intracranial hemorrhage: systematic review and meta-analysis.

OBJECTIVE: Fetal intracranial hemorrhage (ICH) is associated with an increased risk of perinatal mortality and morbidity. Healthcare professionals often find it challenging to counsel parents due to its rarity and diverse presentation. The aim of this systematic review and meta-analysis was to investigate the perinatal outcome of fetuses with ICH. METHODS: MEDLINE, EMBASE, ClinicalTrials.gov and The Cochrane Library databases were searched. Inclusion criteria were studies reporting the outcome of fetuses, newborns and infants diagnosed with ICH. The primary outcome was perinatal death (PND), defined as the sum of intrauterine (IUD) and neonatal death (NND). The secondary outcomes were stillbirth, NND, IUD, termination of pregnancy, need for surgery/shunting at birth, cerebral palsy (defined according to the European Cerebral Palsy Network and classified as diplegia, hemiplegia, quadriplegia, dyskinetic or mixed), neurodevelopmental delay and intact survival. All outcomes were explored in the included fetuses with ICH. A subgroup analysis according to the location of the hemorrhage (intra-axial and extra-axial) was also planned. Meta-analysis of proportions was used to combine data, and pooled proportions and their 95% CI were reported. RESULTS: Sixteen studies (193 fetuses) were included in the meta-analysis. PND occurred in 14.6% (95% CI, 7.3-24.0%) of fetuses with ICH. Among liveborn cases, 27.6% (95% CI, 12.5-45.9%) required shunt placement or surgery after birth and 32.0% (95% CI, 22.2-42.6%) had cerebral palsy. Furthermore, 16.7% (95% CI, 8.4-27.2%) of cases had mild neurodevelopmental delay, while 31.1% (95% CI, 19.0-44.7%) experienced severe adverse neurodevelopmental outcome. Normal neurodevelopmental outcome was reported in 53.6% of fetuses. Subgroup analysis according to the location of ICH showed that PND occurred in 13.3% (95% CI, 5.7-23.4%) of fetuses with intra-axial bleeding and 26.7% (95% CI, 5.3-56.8%) of those with extra-axial bleeding. In fetuses with intra-axial hemorrhage, 25.2% (95% CI, 11.0-42.9%) required shunt placement or surgery after birth and 25.5% (95% CI, 15.3-37.2%) experienced cerebral palsy. In fetuses with intra-axial hemorrhage, mild and severe neurodevelopmental delay was observed in 14.9% (95% CI, 12.0-27.0%) and 32.8% (95% CI, 19.8-47.4%) of cases, respectively, while 53.2% (95% CI, 37.0-69.1%) experienced normal neurodevelopmental outcome. The incidence of mortality and postnatal neurodevelopmental outcome in fetuses with extra-axial hemorrhage could not be estimated reliably due to the small number of cases. CONCLUSIONS: Fetuses with a prenatal diagnosis of ICH are at high risk of perinatal mortality and adverse neurodevelopmental outcome. Postnatal shunt placement or surgery was required in 28% of cases and cerebral palsy was diagnosed in approximately one-third of infants. Due to the rarity of ICH, multicenter prospective registries are warranted to collect high-quality data. © 2021 International Society of Ultrasound in Obstetrics and Gynecology.

First-trimester ultrasound detection of fetal heart anomalies: systematic review and meta-analysis.

OBJECTIVES: To determine the diagnostic accuracy of ultrasound at 11-14 weeks' gestation in the detection of fetal cardiac abnormalities and to evaluate factors that impact the detection rate. METHODS: This was a systematic review of studies evaluating the diagnostic accuracy of ultrasound in the detection of fetal cardiac anomalies at 11-14 weeks' gestation, performed by two independent reviewers. An electronic search of four databases (MEDLINE, EMBASE, Web of Science Core Collection and The Cochrane Library) was conducted for studies published between January 1998 and July 2020. Prospective and retrospective studies evaluating pregnancies at any prior level of risk and in any healthcare setting were eligible for inclusion. The reference standard used was the detection of a cardiac abnormality on postnatal or postmortem examination. Data were extracted from the included studies to populate 2 × 2 tables. Meta-analysis was performed using a random-effects model in order to determine the performance of first-trimester ultrasound in the detection of major cardiac abnormalities overall and of individual types of cardiac abnormality. Data were analyzed separately for high-risk and non-high-risk populations. Preplanned secondary analyses were conducted in order to assess factors that may impact screening performance, including the imaging protocol used for cardiac assessment (including the use of color-flow Doppler), ultrasound modality, year of publication and the index of sonographer suspicion at the time of the scan. Risk of bias and quality assessment were undertaken for all included studies using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool. RESULTS: The electronic search yielded 4108 citations. Following review of titles and abstracts, 223 publications underwent full-text review, of which 63 studies, reporting on 328 262 fetuses, were selected for inclusion in the meta-analysis. In the non-high-risk population (45 studies, 306 872 fetuses), 1445 major cardiac anomalies were identified (prevalence, 0.41% (95% CI, 0.39-0.43%)). Of these, 767 were detected on first-trimester ultrasound examination of the heart and 678 were not detected. First-trimester ultrasound had a pooled sensitivity of 55.80% (95% CI, 45.87-65.50%), specificity of 99.98% (95% CI, 99.97-99.99%) and positive predictive value of 94.85% (95% CI, 91.63-97.32%) in the non-high-risk population. The cases diagnosed in the first trimester represented 63.67% (95% CI, 54.35-72.49%) of all antenatally diagnosed major cardiac abnormalities in the non-high-risk population. In the high-risk population (18 studies, 21 390 fetuses), 480 major cardiac anomalies were identified (prevalence, 1.36% (95% CI, 1.20-1.52%)). Of these, 338 were detected on first-trimester ultrasound examination and 142 were not detected. First-trimester ultrasound had a pooled sensitivity of 67.74% (95% CI, 55.25-79.06%), specificity of 99.75% (95% CI, 99.47-99.92%) and positive predictive value of 94.22% (95% CI, 90.22-97.22%) in the high-risk population. The cases diagnosed in the first trimester represented 79.86% (95% CI, 69.89-88.25%) of all antenatally diagnosed major cardiac abnormalities in the high-risk population. The imaging protocol used for examination was found to have an important impact on screening performance in both populations (P < 0.0001), with a significantly higher detection rate observed in studies using at least one outflow-tract view or color-flow Doppler imaging (both P < 0.0001). Different types of cardiac anomaly were not equally amenable to detection on first-trimester ultrasound. CONCLUSIONS: First-trimester ultrasound examination of the fetal heart allows identification of over half of fetuses affected by major cardiac pathology. Future first-trimester screening programs should follow structured anatomical assessment protocols and consider the introduction of outflow-tract views and color-flow Doppler imaging, as this would improve detection rates of fetal cardiac pathology. © 2021 The Authors. Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology.

Routine first-trimester combined screening for pre-eclampsia: pregnancy-associated plasma protein-A or placental growth factor?

OBJECTIVE: To compare the screening performance of serum pregnancy-associated plasma protein-A (PAPP-A) vs placental growth factor (PlGF) in routine first-trimester combined screening for pre-eclampsia (PE), small-for-gestational age (SGA) at birth and trisomy 21. METHODS: This was a retrospective study nested in pregnancy cohorts undergoing first-trimester combined screening for PE and trisomy 21 using The Fetal Medicine Foundation (FMF) algorithm based on maternal characteristics, nuchal translucency thickness, PAPP-A, free beta-human chorionic gonadotropin, blood pressure and uterine artery Doppler. Women at high risk for preterm PE (≥ 1 in 50) received 150 mg of aspirin per day, underwent serial fetal growth scans at 28 and 36 weeks and were offered elective birth from 40 weeks of gestation. PlGF was quantified retrospectively from stored surplus first-trimester serum samples. The performance of combined first-trimester screening for PE and SGA using maternal history, blood pressure, uterine artery pulsatility index and either PAPP-A or PlGF was calculated. Similarly, the performance of combined first-trimester screening for trisomy 21 was calculated using either PAPP-A or PlGF in addition to maternal age, nuchal translucency thickness and free beta-human chorionic gonadotropin. RESULTS: Maternal serum PAPP-A was assayed in 1094 women, including 82 with PE, 111 with SGA (birth weight < 10th centile), 53 with both PE and SGA and 94 with fetal trisomy 21. PlGF levels were obtained retrospectively from 1066/1094 women. Median serum PlGF multiples of the median was significantly lower in pregnancies with PE (1.0 (interquartile range (IQR), 0.8-1.4); P < 0.01), SGA (1.0 (IQR, 0.8-1.3); P < 0.001) and trisomy 21 (0.6 (IQR, 0.5-0.9); P < 0.0001) compared to in controls (1.2 (IQR, 0.9-1.5)). There was no significant difference in the performance of first-trimester screening using PAPP-A vs PlGF for either preterm PE (area under the receiver-operating-characteristics curve (AUC), 0.78 vs 0.79; P = 0.55) or term PE (AUC, 0.74 vs 0.74; P = 0.60). These findings persisted even after correction for the effect of targeted aspirin use on the prevalence of PE. Similarly, there were no significant differences in sensitivity and specificity of combined screening for SGA or trisomy 21 when using PAPP-A vs PlGF. CONCLUSIONS: Using either PlGF or PAPP-A in routine first-trimester combined screening based on maternal characteristics, blood pressure and uterine artery Doppler does not make a significant clinical difference to the detection of PE or SGA. Depending on the setting, biomarkers should be chosen to achieve a good compromise between performance and measurement requirements. This pragmatic clinical-effectiveness study suggests that combined screening for PE can be implemented successfully in a public healthcare setting without changing current protocols for the assessment of PAPP-A in the first trimester. © 2021 The Authors. Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology.

Influence of maternal body mass index on interobserver variability of fetal ultrasound biometry and amniotic-fluid assessment in late pregnancy.

OBJECTIVE: To determine the interobserver reproducibility of fetal ultrasound biometric and amniotic-fluid measurements in the third trimester of pregnancy, according to maternal body mass index (BMI) category. METHODS: This was a prospective cohort study of women with a singleton gestation beyond 34 weeks, recruited into four groups according to BMI category: normal (18.0-24.9 kg/m2 ), overweight (25.0-29.9 kg/m2) , obese (30.0-39.9 kg/m2 ) and morbidly obese (≥ 40 kg/m2 ). Multiple pregnancies, women with diabetes and pregnancies with a fetal growth, structural or genetic abnormality were excluded. In each woman, fetal biometric (biparietal diameter (BPD), head circumference, abdominal circumference (AC), femur length (FL) and estimated fetal weight) and amniotic-fluid (amniotic-fluid index (AFI) and maximum vertical pocket (MVP)) measurements were obtained by two experienced sonographers or physicians, blinded to gestational age and each other's measurements. Differences in measurements between observers were expressed as gestational age-specific Z-scores. The interobserver intraclass correlation coefficient (ICC) and Cronbach's reliability coefficient (CRC) were calculated. Bland-Altman analysis was used to assess the degree of reproducibility. RESULTS: In total, 110 women were enrolled prospectively (including 1320 measurements obtained by 17 sonographers or physicians). Twenty (18.2%) women had normal BMI, 30 (27.3%) women were overweight, 30 (27.3%) women were obese and 30 (27.3%) women were morbidly obese. Except for AFI (ICC, 0.65; CRC, 0.78) and MVP (ICC, 0.49; CRC, 0.66), all parameters had a very high level of interobserver reproducibility (ICC, 0.72-0.87; CRC, 0.84-0.93). When assessing reproducibility according to BMI category, BPD measurements had a very high level of reproducibility (ICC ≥ 0.85; CRC > 0.90) in all groups. The reproducibility of AC and FL measurements increased with increasing BMI, while the reproducibility of MVP measurements decreased. Among the biometric parameters, the difference between the BMI categories in measurement-difference Z-score was significant only for FL. Interobserver differences for biometric measurements fell within the 95% limits of agreement. CONCLUSION: Obesity does not seem to impact negatively on the reproducibility of ultrasound measurements of fetal biometric parameters when undertaken by experienced sonographers or physicians who commonly assess overweight, obese and morbidly obese women. © 2021 International Society of Ultrasound in Obstetrics and Gynecology.

How often do we identify fetal abnormalities during routine third-trimester ultrasound? A systematic review and meta-analysis.

BACKGROUND: Routine third-trimester ultrasound is frequently offered to pregnant women to identify fetuses with abnormal growth. Infrequently, a congenital anomaly is incidentally detected. OBJECTIVE: To establish the prevalence and type of fetal anomalies detected during routine third-trimester scans using a systematic review and meta-analysis. SEARCH STRATEGY: Electronic databases (MEDLINE, Embase and the Cochrane library) from inception until August 2019. SELECTION CRITERIA: Population-based studies (randomised control trials, prospective and retrospective cohorts) reporting abnormalities detected at the routine third-trimester ultrasound performed in unselected populations with prior screening. Case reports, case series, case-control studies and reviews without original data were excluded. DATA COLLECTION AND ANALYSIS: Prevalence and type of anomalies detected in the third trimester. We calculated pooled prevalence as the number of anomalies per 1000 scans with 95% confidence intervals. Publication bias was assessed. MAIN RESULTS: The literature search identified 9594 citations: 13 studies were eligible representing 141 717 women; 643 were diagnosed with an unexpected abnormality. The pooled prevalence of a new abnormality diagnosed was 3.68 per 1000 women scanned (95% CI 2.72-4.78). The largest groups of abnormalities were urogenital (55%), central nervous system abnormalities (18%) and cardiac abnormalities (14%). CONCLUSION: Combining data from 13 studies and over 140 000 women, we show that during routine third-trimester ultrasound, an incidental fetal anomaly will be found in about 1 in 300 scanned women. This information should be taken into account when taking consent from women for third-trimester ultrasound and when designing and assessing cost of third-trimester ultrasound screening programmes. TWEETABLE ABSTRACT: One in 300 women attending a third-trimester scan will have a finding of a fetal abnormality.

Fetal cerebellar growth and Sylvian fissure maturation: international standards from Fetal Growth Longitudinal Study of INTERGROWTH-21st Project.

OBJECTIVE: To construct international ultrasound-based standards for fetal cerebellar growth and Sylvian fissure maturation. METHODS: Healthy, well nourished pregnant women, enrolled at < 14 weeks' gestation in the Fetal Growth Longitudinal Study (FGLS) of INTERGROWTH-21st , an international multicenter, population-based project, underwent serial three-dimensional (3D) fetal ultrasound scans every 5 ± 1 weeks until delivery in study sites located in Brazil, India, Italy, Kenya and the UK. In the present analysis, only those fetuses that underwent developmental assessment at 2 years of age were included. We measured the transcerebellar diameter and assessed Sylvian fissure maturation using two-dimensional ultrasound images extracted from available 3D fetal head volumes. The appropriateness of pooling data from the five sites was assessed using variance component analysis and standardized site differences. For each Sylvian fissure maturation score (left or right side), mean gestational age and 95% CI were calculated. Transcerebellar diameter was modeled using fractional polynomial regression, and goodness of fit was assessed. RESULTS: Of those children in the original FGLS cohort who had developmental assessment at 2 years of age, 1130 also had an available 3D ultrasound fetal head volume. The sociodemographic characteristics and pregnancy/perinatal outcomes of the study sample confirmed the health and low-risk status of the population studied. In addition, the fetuses had low morbidity and adequate growth and development at 2 years of age. In total, 3016 and 2359 individual volumes were available for transcerebellar-diameter and Sylvian-fissure analysis, respectively. Variance component analysis and standardized site differences showed that the five study populations were sufficiently similar on the basis of predefined criteria for the data to be pooled to produce international standards. A second-degree fractional polynomial provided the best fit for modeling transcerebellar diameter; we then estimated gestational-age-specific 3rd , 50th and 97th smoothed centiles. Goodness-of-fit analysis comparing empirical centiles with smoothed centile curves showed good agreement. The Sylvian fissure increased in maturation with advancing gestation, with complete overlap of the mean gestational age and 95% CIs between the sexes for each development score. No differences in Sylvian fissure maturation between the right and left hemispheres were observed. CONCLUSION: We present, for the first time, international standards for fetal cerebellar growth and Sylvian fissure maturation throughout pregnancy based on a healthy fetal population that exhibited adequate growth and development at 2 years of age. © 2020 The Authors. Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology.

Discovering Salient Anatomical Landmarks by Predicting Human Gaze.

Anatomical landmarks are a crucial prerequisite for many medical imaging tasks. Usually, the set of landmarks for a given task is predefined by experts. The landmark locations for a given image are then annotated manually or via machine learning methods trained on manual annotations. In this paper, in contrast, we present a method to automatically discover and localize anatomical landmarks in medical images. Specifically, we consider landmarks that attract the visual attention of humans, which we term visually salient landmarks. We illustrate the method for fetal neurosonographic images. First, full-length clinical fetal ultrasound scans are recorded with live sonographer gaze-tracking. Next, a convolutional neural network (CNN) is trained to predict the gaze point distribution (saliency map) of the sonographers on scan video frames. The CNN is then used to predict saliency maps of unseen fetal neurosonographic images, and the landmarks are extracted as the local maxima of these saliency maps. Finally, the landmarks are matched across images by clustering the landmark CNN features. We show that the discovered landmarks can be used within affine image registration, with average landmark alignment errors between 4.1% and 10.9% of the fetal head long axis length.

Introduction to artificial intelligence in ultrasound imaging in obstetrics and gynecology.

Artificial intelligence (AI) uses data and algorithms to aim to draw conclusions that are as good as, or even better than, those drawn by humans. AI is already part of our daily life; it is behind face recognition technology, speech recognition in virtual assistants (such as Amazon Alexa, Apple's Siri, Google Assistant and Microsoft Cortana) and self-driving cars. AI software has been able to beat world champions in chess, Go and recently even Poker. Relevant to our community, it is a prominent source of innovation in healthcare, already helping to develop new drugs, support clinical decisions and provide quality assurance in radiology. The list of medical image-analysis AI applications with USA Food and Drug Administration or European Union (soon to fall under European Union Medical Device Regulation) approval is growing rapidly and covers diverse clinical needs, such as detection of arrhythmia using a smartwatch or automatic triage of critical imaging studies to the top of the radiologist's worklist. Deep learning, a leading tool of AI, performs particularly well in image pattern recognition and, therefore, can be of great benefit to doctors who rely heavily on images, such as sonologists, radiographers and pathologists. Although obstetric and gynecological ultrasound are two of the most commonly performed imaging studies, AI has had little impact on this field so far. Nevertheless, there is huge potential for AI to assist in repetitive ultrasound tasks, such as automatically identifying good-quality acquisitions and providing instant quality assurance. For this potential to thrive, interdisciplinary communication between AI developers and ultrasound professionals is necessary. In this article, we explore the fundamentals of medical imaging AI, from theory to applicability, and introduce some key terms to medical professionals in the field of ultrasound. We believe that wider knowledge of AI will help accelerate its integration into healthcare. © 2020 The Authors. Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of the International Society of Ultrasound in Obstetrics and Gynecology.

Reduction in twin stillbirth following implementation of NICE guidance.

OBJECTIVE: There has been an unprecedented fall in the rate of stillbirth in twin pregnancy in the UK. It is contested whether implementation of the National Institute for Health and Care Excellence (NICE) guidance on the antenatal management of uncomplicated twin pregnancies has contributed to this change. The aim of this study was to investigate whether the implementation of NICE guidance was associated with a reduction in the rate of stillbirth in twin pregnancies delivered in a large UK hospital. METHODS: This was a retrospective cohort study including all twin pregnancies delivered at St George's Hospital, London, UK, between 2000 and 2018. Data were analyzed according to two time periods: before implementation of the NICE guidance on twins (before June 2013; pre-NICE) and after its implementation (after June 2013; post-NICE). The exclusion criteria were higher-order multiple gestations, pregnancies of unknown chorionicity, pregnancies complicated by miscarriage, those that underwent termination and those diagnosed with vanishing twin. The main outcome was stillbirth. Other outcomes included neonatal death (NND), admission to the neonatal intensive care unit (NICU) and emergency Cesarean section. We planned a priori a sensitivity analysis according to chorionicity. The chi-square test and Mann-Whitney U-test were used to compare outcomes between the study groups. RESULTS: We included in the analysis 1666 twin pregnancies (3332 fetuses), of which 1114 pregnancies (2228 fetuses) were delivered before and 552 pregnancies (1104 fetuses) after June 2013. Of those, 1299 were dichorionic and 354 were monochorionic diamniotic. The incidence of stillbirth was significantly lower in the post-NICE than in the pre-NICE group (3.6 per 1000 births vs 13.5 per 1000 births; P = 0.008). The reduction in stillbirth rate was from 8.5 to 3.6 per 1000 births (P = 0.161) in dichorionic and from 33.6 to 3.8 per 1000 births (P = 0.011) in monochorionic diamniotic twin pregnancies. There was no significant difference in the rates of NND (P = 0.625), NICU admission (P = 0.506) or emergency Cesarean section (P = 0.820) between the two groups. The median gestational age at delivery was significantly lower in the post-NICE than in the pre-NICE group (median 36.3 vs 36.9 weeks; P < 0.001), as a consequence of a significant increase in preterm birth between 34 and 37 weeks' gestation (39.3% vs 27.0%; P = 0.002), but not before 34 weeks (P = 0.473). CONCLUSIONS: A significant reduction of > 70% in the stillbirth rate in twin pregnancies was noted after implementation of the NICE guidance. This reduction was statistically significant in monochorionic, but not dichorionic, twin pregnancies. The improvement in twin pregnancy outcome was achieved without a concomitant increase in NND, admission to the NICU or emergency Cesarean section. © 2020 The Authors. Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of the International Society of Ultrasound in Obstetrics and Gynecology.