Ensemble learning for fetal ultrasound and maternal-fetal data to predict mode of delivery after labor induction.

Providing adequate counseling on mode of delivery after induction of labor (IOL) is of utmost importance. Various AI algorithms have been developed for this purpose, but rely on maternal-fetal data, not including ultrasound (US) imaging. We used retrospectively collected clinical data from 808 subjects submitted to IOL, totaling 2024 US images, to train AI models to predict vaginal delivery (VD) and cesarean section (CS) outcomes after IOL. The best overall model used only clinical data (F1-score: 0.736; positive predictive value (PPV): 0.734). The imaging models employed fetal head, abdomen and femur US images, showing limited discriminative results. The best model used femur images (F1-score: 0.594; PPV: 0.580). Consequently, we constructed ensemble models to test whether US imaging could enhance the clinical data model. The best ensemble model included clinical data and US femur images (F1-score: 0.689; PPV: 0.693), presenting a false positive and false negative interesting trade-off. The model accurately predicted CS on 4 additional cases, despite misclassifying 20 additional VD, resulting in a 6.0% decrease in average accuracy compared to the clinical data model. Hence, integrating US imaging into the latter model can be a new development in assisting mode of delivery counseling.

How to Perform Fetal MR Imaging.

This article provides the readers with practical guidance on how to perform fetal MR imaging, including technical considerations such as scanner field strength and use of appropriate radiofrequency receive coils, and summarizes the role, strengths, and limitations of the various MR imaging sequences. The authors review the various factors to consider in scan preparation, including study indication, timing, maternal preparation, and the creation of an institutional fetal imaging protocol. Additional factors that go into scan optimization during acquisition including prioritizing maternal comfort and ways to troubleshoot various artifacts that maybe encountered in fetal imaging are discussed.

Evaluation of embryo-foetal biometry and its correlation with parturition date in Toy Poodle bitches.

Estimating the parturition date in dogs is challenging due to their reproductive peculiarities that. Ultrasonographic examination serves as a tool for studying embryo/foetal biometry and estimating the time of parturition by measuring foetal and extra-foetal structures. However, due to reproductive differences among various dog breeds, such estimates may have a non-significant pattern, representing inaccuracies in the estimated date of birth. This study aimed to monitor pregnant Toy Poodle bitches and establish relationships between ultrasonographically measured foetal and extra-foetal dimensions and the remaining time until parturition. Eighteen pregnant Toy Poodle bitches were subjected to weekly ultrasonographic evaluations and measurements of the inner chorionic cavity diameter, craniocaudal length (CCL), biparietal diameter (BPD), diameter of the deep portion of diencephalo-telencephalic vesicle (DPTV), abdominal diameter, thorax diameter (TXD), placental thickness and the renal diameter (REND). These parameters were retrospectively correlated with the date of parturition and linear regressions were established between gestational measurements and days before parturition (DBP). All analyses were conducted using the Statistical Package for Social Sciences (IBM® SPSS®) program at a 5% significance level. The foetal measurements that showed a high correlation (r) and reliability (R2) with DBP were BPD [(DBP = [15.538 × BPD] - 39.756), r = .97 and R2 = .93], TXD [(DBP = [8.933 × TXD] - 32.487), r = .94 and R2 = .89], DPTV [(DBP = [34.580 × DPTV] - 39.403), r = .93 and R2 = .86] and REND [(DBP = [13.735 × REND] - 28.937), r = .91 and R2 = .82]. This statistically validates the application of these specific formulas to estimate the parturition date in Toy Poodle bitches.

First-trimester 3D fetal neurosonography: five standardised views.

BACKGROUND: There are several international guidelines for foetal anomalies scanning at 11-14 weeks' gestation. The aim of this study is to present our first-trimester specialist neurosonography protocol with examples of pathology in order to develop a systematic approach to evaluating the first-trimester foetal brain. METHODS: Women undergoing a first-trimester foetal medicine ultrasound scan between 2010 and 2020 for multiple indications underwent neurosonography according to a set protocol. 3D transvaginal brain examination was performed in all cases (2000 pregnancies scanned). We retrospectively reviewed all imaging to develop this protocol. RESULTS: We propose that the following five axial-plane parallel views should be obtained when performing neurosonography in the first trimester, moving from cranial to caudal: 1. Lateral ventricles; 2. Third ventricle; 3. Thalamus and mesencephalon; 4. Cerebellum; 5. Fourth ventricle. Examples of these images and abnormalities that can be seen in each plane are given. CONCLUSIONS: We have presented a specialist protocol for systematically assessing the foetal brain in the first trimester and given examples of pathology which may be seen in each plane. Further work is needed to prospectively assess detection rates of major abnormalities using this protocol and assess the reproducibility and learning curve of this technique.

This article suggests a way in which specialists scanning babies at 11­14 weeks of pregnancy can check the brain in a structured way. This involves looking at the brain at five levels or planes to view the developing structures. The suggested scan protocol is similar to images produced of the brain and heart at the second trimester (20 week) scan. We hope that specialists will find it useful to check the brain in this way if there are concerns raised at the dating (12 week) scan, and that this will lead to earlier detection of brain abnormalities or differences.

Prognosis of remaining fetus in twin pregnancy after demise of one fetus according to its location.

BACKGROUND: To investigate the prognosis of the remaining fetus in twin pregnancy after experiencing one fetal demise in the first trimester according to the location of the demised fetus. METHODS: This was a retrospective study of twin pregnancies with one fetal demise after the first trimester (14 weeks of gestation) delivered between September 2004 and September 2022. The study population was divided into two groups based on the location of the demised fetus as determined by the last recorded ultrasonography results: Group 1 included twin pregnancies where the presenting fetus was demised (n = 36) and Group 2 included twin pregnancies where the non-presenting fetus was demised (n = 44). The obstetric and neonatal outcomes were also reviewed. RESULTS: A total of 80 pregnant women were included. The median gestational age for the diagnosis of fetal demise was 24.1 weeks. The gestational age of the demised fetus was not different between Groups 1 and 2; however, the gestational age of the remaining fetus at delivery was significantly earlier in Group 1 than it was in Group 2 (33.8 vs. 37.3 weeks, P = .004). The rate of preterm birth before 28 weeks was almost five times higher in Group 1 than in Group 2 (22.2% vs. 4.5%, P = .037). Regression analysis demonstrated significant differences between Groups 1 and 2. Respiratory distress syndrome, bronchopulmonary dysplasia, patent ductus arteriosus, retinopathy of prematurity, and jaundice were more common in Group 1 than in Group 2; however, the association was not significant after adjusting for gestational age at delivery. CONCLUSIONS: When the presenting fetus is demised in a twin pregnancy, the remaining fetus tends to be delivered earlier than when the non-presenting fetus is demised.

Quantitative study of the ossification centers of the body of sphenoid bone in the human fetus.

The aim of the present study was to examine the growth dynamics of the two ossification centers of the body of sphenoid bone in the human fetus, based on their linear, planar and volumetric parameters. The examinations were carried out on 37 human fetuses of both sexes aged 18-30 weeks of gestation, which had been preserved in 10% neutral formalin solution. Using CT, digital image analysis software, 3D reconstruction and statistical methods, we evaluated the size of the presphenoid and postsphenoid ossification centers. The presphenoid ossification center grew proportionately in sagittal diameter, projection surface area and volume, and logarithmically in transverse diameter. The postsphenoid ossification center increased logarithmically in sagittal diameter, transverse diameter and projection surface area, while its volumetric growth followed proportionately. The numerical findings of the presphenoid and postsphenoid ossification centers may be considered age-specific reference values of potential relevance in monitoring the normal fetal growth and screening for congenital disorders in the fetus. The obtained results may contribute to a better understanding of the growing fetal skeleton, bringing new numerical information regarding its diagnosis and development.

3D atlas of the human fetal chondrocranium in the middle trimester.

The chondrocranium provides the key initial support for the fetal brain, jaws and cranial sensory organs in all vertebrates. The patterns of shaping and growth of the chondrocranium set up species-specific development of the entire craniofacial complex. The 3D development of chondrocranium have been studied primarily in animal model organisms, such as mice or zebrafish. In comparison, very little is known about the full 3D human chondrocranium, except from drawings made by anatomists many decades ago. The knowledge of human-specific aspects of chondrocranial development are essential for understanding congenital craniofacial defects and human evolution. Here advanced microCT scanning was used that includes contrast enhancement to generate the first 3D atlas of the human fetal chondrocranium during the middle trimester (13 to 19 weeks). In addition, since cartilage and bone are both visible with the techniques used, the  endochondral ossification of cranial base was mapped since this region is so critical for brain and jaw growth. The human 3D models are published as a scientific resource for human development.

Advances in Fetal Brain Imaging.

Over the last 20 years, there have been remarkable developments in fetal brain MR imaging analysis methods. This article delves into the specifics of structural imaging, diffusion imaging, functional MR imaging, and spectroscopy, highlighting the latest advancements in motion correction, fetal brain development atlases, and the challenges and innovations. Furthermore, this article explores the clinical applications of these advanced imaging techniques in comprehending and diagnosing fetal brain development and abnormalities.

Technical note: Capturing shape-Linear measurements and geometric morphometrics of the immature femora.

OBJECTIVES: Growth and developmental studies have been a prominent theme in bioarchaeology. These works traditionally focus on metric measurements of long bone length and age-at-death or cross-sectional geometric studies with the use of computed tomography scans for questions on growth and mobility. However, teasing apart aspects of size and shape have been difficult due to the cylindrical nature of immature long bones. This research investigates the methodological use of surface geometries from linear measurements and geometric morphometric methods (GMM) to answer questions on mobility and allometry during childhood. MATERIALS AND METHODS: Left femora were selected from 42 individuals ranging from fetal to 12 years of age from medieval St Gregory's Priory, Canterbury, UK. Femora were digitized with structured-light-scanning for auto3dgm analysis and measurements were obtained from physical caliper measurements. Individuals were put into age groups based on biomechanical milestones during this age range. RESULTS: Ratio and GMM confirm hypotheses of allometry and biomechanical milestones. Geometric morphometrics, however, detects more subtle differences in mobility at each age group. DISCUSSION: The findings of this preliminary study support the potential use of GMM of immature femora, while indicating that the extent in range of mobility that can occur varies at different biological milestones.

Evaluation of twin fetal exposure to radiofrequency field during magnetic resonance imaging.

Fetal development is essential to the human lifespan. As more and more multifetal gestations have been reported recently, clinical diagnosis using magnetic resonance imaging (MRI), which introduced radiofrequency (RF) exposure, raised public concerns. The present study developed two whole-body pregnant models of 31 and 32 gestational weeks (GWs) with twin fetuses and explored RF exposure by 1.5 and 3.0 T MRI. Differences in the relative position of the fetus and changes in fetal weight can cause differences in fetal peak local specific absorption rate averaged over 10 g tissue (pSAR10g). Variation of pSAR10g due to different fetal positions can be ~35%. Numerically, twin and singleton fetal pSAR10g results were not significantly different, however twin results exceeded the limit in some cases (e.g. fetuses of 31 GW at 1.5 T), which indicated the necessity for further research employing anatomically correct twin-fetal models coming from various GWs and particular sequence to be applied.

Estimated Fetal Radiation Dose From 18 F-FDG PET/CT During the Second and Third Trimesters of Pregnancy.

PURPOSE: Data published in the literature concerning the doses received by fetuses exposed to a 18 F-FDG PET are reassuring but were obtained from small and heterogeneous cohorts, and very few data are available concerning the fetal dose received after exposure to both PET and CT. The present study aimed to estimate the fetal dose received following a PET/CT exposure using methods that include anthropomorphic phantoms of pregnant women applied on a large cohort. PATIENTS AND METHODS: This retrospective multicenter study included 18 pregnant patients in the second and third trimesters. For PET exposure, the fetal volume and mean concentration of radioactivity in the fetus were measured by manually drawing regions of interest. Those data, combined with the time-integrated activities of the fetus and the mother's organs, were entered into the OLINDA/EXM software 2.0 to assess the fetal dose due to PET exposure. To estimate the fetal dose received due to CT exposure, 2 softwares were used: CT-Expo (based on geometric phantom models of nonpregnant patients) and VirtualDose (using pregnant patient phantoms). RESULTS: The fetal dose exposure for PET/CT examination in the second trimester ranged from 5.7 to 15.8 mGy using CT-Expo (mean, 11.6 mGy) and from 5.1 to 11.6 mGy using VirtualDose (mean, 8.6 mGy). In the third trimester, it ranged from 7.9 to 16.6 mGy using CT-Expo (mean, 10.7 mGy) and from 6.1 to 10.7 mGy using VirtualDose (mean, 7.6 mGy). CONCLUSIONS: The estimated fetal doses were in the same range of those previously published and are well below the threshold for deterministic effects. Pregnancy does not constitute an absolute contraindication for a clinically justified hybrid 18 F-FDG PET/CT.

Investigation on ultrasound images for detection of fetal congenital heart defects.

Congenital heart defects (CHD) are one of the serious problems that arise during pregnancy. Early CHD detection reduces death rates and morbidity but is hampered by the relatively low detection rates (i.e., 60%) of current screening technology. The detection rate could be increased by supplementing ultrasound imaging with fetal ultrasound image evaluation (FUSI) using deep learning techniques. As a result, the non-invasive foetal ultrasound image has clear potential in the diagnosis of CHD and should be considered in addition to foetal echocardiography. This review paper highlights cutting-edge technologies for detecting CHD using ultrasound images, which involve pre-processing, localization, segmentation, and classification. Existing technique of preprocessing includes spatial domain filter, non-linear mean filter, transform domain filter, and denoising methods based on Convolutional Neural Network (CNN); segmentation includes thresholding-based techniques, region growing-based techniques, edge detection techniques, Artificial Neural Network (ANN) based segmentation methods, non-deep learning approaches and deep learning approaches. The paper also suggests future research directions for improving current methodologies.

RTSeg-net: A lightweight network for real-time segmentation of fetal head and pubic symphysis from intrapartum ultrasound images.

The segmentation of the fetal head (FH) and pubic symphysis (PS) from intrapartum ultrasound images plays a pivotal role in monitoring labor progression and informing crucial clinical decisions. Achieving real-time segmentation with high accuracy on systems with limited hardware capabilities presents significant challenges. To address these challenges, we propose the real-time segmentation network (RTSeg-Net), a groundbreaking lightweight deep learning model that incorporates innovative distribution shifting convolutional blocks, tokenized multilayer perceptron blocks, and efficient feature fusion blocks. Designed for optimal computational efficiency, RTSeg-Net minimizes resource demand while significantly enhancing segmentation performance. Our comprehensive evaluation on two distinct intrapartum ultrasound image datasets reveals that RTSeg-Net achieves segmentation accuracy on par with more complex state-of-the-art networks, utilizing merely 1.86 M parameters-just 6 % of their hyperparameters-and operating seven times faster, achieving a remarkable rate of 31.13 frames per second on a Jetson Nano, a device known for its limited computing capacity. These achievements underscore RTSeg-Net's potential to provide accurate, real-time segmentation on low-power devices, broadening the scope for its application across various stages of labor. By facilitating real-time, accurate ultrasound image analysis on portable, low-cost devices, RTSeg-Net promises to revolutionize intrapartum monitoring, making sophisticated diagnostic tools accessible to a wider range of healthcare settings.

Sexual dimorphism of the human fetal pelvis exists at the onset of primary ossification.

Human adolescent and adult skeletons exhibit sexual dimorphism in the pelvis. However, the degree of sexual dimorphism of the human pelvis during prenatal development remains unclear. Here, we performed high-resolution magnetic resonance imaging-assisted pelvimetry on 72 human fetuses (males [M]: females [F], 34:38; 21 sites) with crown-rump lengths (CRL) of 50-225 mm (the onset of primary ossification). We used multiple regression analysis to examine sexual dimorphism with CRL as a covariate. Females exhibit significantly smaller pelvic inlet anteroposterior diameters (least squares mean, [F] 8.4 mm vs. [M] 8.8 mm, P = 0.036), larger subpubic angle ([F] 68.1° vs. [M] 64.0°, P = 0.034), and larger distance between the ischial spines relative to the transverse diameters of the greater pelvis than males. Furthermore, the sacral measurements indicate significant sex-CRL interactions. Our study suggests that sexual dimorphism of the human fetal pelvis is already apparent at the onset of primary ossification.

Deep learning microstructure estimation of developing brains from diffusion MRI: A newborn and fetal study.

Diffusion-weighted magnetic resonance imaging (dMRI) is widely used to assess the brain white matter. Fiber orientation distribution functions (FODs) are a common way of representing the orientation and density of white matter fibers. However, with standard FOD computation methods, accurate estimation requires a large number of measurements that usually cannot be acquired for newborns and fetuses. We propose to overcome this limitation by using a deep learning method to map as few as six diffusion-weighted measurements to the target FOD. To train the model, we use the FODs computed using multi-shell high angular resolution measurements as target. Extensive quantitative evaluations show that the new deep learning method, using significantly fewer measurements, achieves comparable or superior results than standard methods such as Constrained Spherical Deconvolution and two state-of-the-art deep learning methods. For voxels with one and two fibers, respectively, our method shows an agreement rate in terms of the number of fibers of 77.5% and 22.2%, which is 3% and 5.4% higher than other deep learning methods, and an angular error of 10° and 20°, which is 6° and 5° lower than other deep learning methods. To determine baselines for assessing the performance of our method, we compute agreement metrics using densely sampled newborn data. Moreover, we demonstrate the generalizability of the new deep learning method across scanners, acquisition protocols, and anatomy on two clinical external datasets of newborns and fetuses. We validate fetal FODs, successfully estimated for the first time with deep learning, using post-mortem histological data. Our results show the advantage of deep learning in computing the fiber orientation density for the developing brain from in-vivo dMRI measurements that are often very limited due to constrained acquisition times. Our findings also highlight the intrinsic limitations of dMRI for probing the developing brain microstructure.

Statistical Characterisation of Fetal Anatomy in Simple Obstetric Ultrasound Video Sweeps.

OBJECTIVE: We present a statistical characterisation of fetal anatomies in obstetric ultrasound video sweeps where the transducer follows a fixed trajectory on the maternal abdomen. METHODS: Large-scale, frame-level manual annotations of fetal anatomies (head, spine, abdomen, pelvis, femur) were used to compute common frame-level anatomy detection patterns expected for breech, cephalic, and transverse fetal presentations, with respect to video sweep paths. The patterns, termed statistical heatmaps, quantify the expected anatomies seen in a simple obstetric ultrasound video sweep protocol. In this study, a total of 760 unique manual annotations from 365 unique pregnancies were used. RESULTS: We provide a qualitative interpretation of the heatmaps assessing the transducer sweep paths with respect to different fetal presentations and suggest ways in which the heatmaps can be applied in computational research (e.g., as a machine learning prior). CONCLUSION: The heatmap parameters are freely available to other researchers (https://github.com/agleed/calopus_statistical_heatmaps).

PSFHS: Intrapartum ultrasound image dataset for AI-based segmentation of pubic symphysis and fetal head.

During the process of labor, the intrapartum transperineal ultrasound examination serves as a valuable tool, allowing direct observation of the relative positional relationship between the pubic symphysis and fetal head (PSFH). Accurate assessment of fetal head descent and the prediction of the most suitable mode of delivery heavily rely on this relationship. However, achieving an objective and quantitative interpretation of the ultrasound images necessitates precise PSFH segmentation (PSFHS), a task that is both time-consuming and demanding. Integrating the potential of artificial intelligence (AI) in the field of medical ultrasound image segmentation, the development and evaluation of AI-based models rely significantly on access to comprehensive and meticulously annotated datasets. Unfortunately, publicly accessible datasets tailored for PSFHS are notably scarce. Bridging this critical gap, we introduce a PSFHS dataset comprising 1358 images, meticulously annotated at the pixel level. The annotation process adhered to standardized protocols and involved collaboration among medical experts. Remarkably, this dataset stands as the most expansive and comprehensive resource for PSFHS to date.

Estimation of maternal and foetal risk of radiation-induced cancer from a survey of computed tomography pulmonary angiography and ventilation/perfusion lung scanning for diagnosing pulmonary embolism during pregnancy.

INTRODUCTION: While there are many papers on maternal and foetal radiation doses from computed tomography pulmonary angiography (CTPA) and ventilation/perfusion (V/Q) lung scanning examinations for diagnosing pulmonary embolism in pregnant patients, few have used clinical data to examine the patient lifetime attributable risk (LAR) of different cancer types. This paper aims to estimate the cancer risk from maternal radiation doses from CTPA and V/Q examinations and associated foetal doses. METHODS: Dosimetric data were determined for 267 pregnant patients who received CTPA and/or V/Q examinations over 8 years. Organ and foetal doses were determined using software allowing patient size variations for CTPA and using two different activity-to-organ dose conversion methods for V/Q scans. The LAR of cancer incidence was estimated using International Commission on Radiological Protection (ICRP) modelling including estimates of detriment. RESULTS: Estimated total cancer incidence was 23 and 22 cases per 100,000 for CTPA and V/Q examinations, respectively, with detriment estimates of 18 and 20 cases. Cancer incidence was evenly divided between lung and breast cancer for CTPA with lung cancer being 80% of all cancer for V/Q. The median foetal doses were 0.03 mSv for CTPA and 0.29 mSv for V/Q. Significant differences in estimated foetal dose for V/Q scans were obtained by the two different methods used. The differences in dose between the modes of CTPA scan acquisition highlight the importance of optimisation. CONCLUSION: Maternal cancer incidence and detriment were remarkably similar for each examination. Optimisation of examinations is critical for low-dose outcomes, particularly for CTPA examination.