Sensorless volumetric reconstruction of fetal brain freehand ultrasound scans with deep implicit representation.

Three-dimensional (3D) ultrasound imaging has contributed to our understanding of fetal developmental processes by providing rich contextual information of the inherently 3D anatomies. However, its use is limited in clinical settings, due to the high purchasing costs and limited diagnostic practicality. Freehand 2D ultrasound imaging, in contrast, is routinely used in standard obstetric exams, but inherently lacks a 3D representation of the anatomies, which limits its potential for more advanced assessment. Such full representations are challenging to recover even with external tracking devices due to internal fetal movement which is independent from the operator-led trajectory of the probe. Capitalizing on the flexibility offered by freehand 2D ultrasound acquisition, we propose ImplicitVol to reconstruct 3D volumes from non-sensor-tracked 2D ultrasound sweeps. Conventionally, reconstructions are performed on a discrete voxel grid. We, however, employ a deep neural network to represent, for the first time, the reconstructed volume as an implicit function. Specifically, ImplicitVol takes a set of 2D images as input, predicts their locations in 3D space, jointly refines the inferred locations, and learns a full volumetric reconstruction. When testing natively-acquired and volume-sampled 2D ultrasound video sequences collected from different manufacturers, the 3D volumes reconstructed by ImplicitVol show significantly better visual and semantic quality than the existing interpolation-based reconstruction approaches. The inherent continuity of implicit representation also enables ImplicitVol to reconstruct the volume to arbitrarily high resolutions. As formulated, ImplicitVol has the potential to integrate seamlessly into the clinical workflow, while providing richer information for diagnosis and evaluation of the developing brain.

Birthweight of children with isolated congenital heart disease-A sibling analysis study.

OBJECTIVE: Congenital heart disease (CHD) is associated with decreased birthweight (BW) compared to population-based references. The aim of this study was to compare the BW of isolated CHD cases to their siblings, thus controlling for unknown and unmeasured confounders within the family. METHODS: All isolated CHD cases in the Leiden University Medical Center were included (2002-2019). Generalized estimated equation models were constructed to compare BW z scores of CHD neonates with their siblings. Cases were clustered to minor or severe CHD and stratified according to the aortic flow and oxygenation to the brain. RESULTS: The overall BW z score of siblings was 0.032 (n = 471). The BW z score was significantly lower in CHD cases (n = 291) compared to their siblings (-0.20, p = 0.005). The results were consistent in the subgroup analysis of severe and minor CHD (BW z score difference -0.20 and -0.10), but did not differ significantly (p = 0.63). Stratified analysis regarding flow and oxygenation showed no BW difference between the groups (p = 0.1). CONCLUSION: Isolated CHD cases display a significantly lower BW z score compared to their siblings. As the siblings of these CHD cases show a BW distribution similar to the general population, this suggests that shared environmental and maternal influences between siblings do not explain the difference in BW.

Birthweight and isolated congenital heart defects - A systematic review and meta-analysis.

BACKGROUND: Birthweight (BW) is an important prognostic factor in newborns with congenital heart defects (CHD). OBJECTIVES: To give an overview of the literature on BW z-score in children with isolated CHD. SEARCH STRATEGY: A systematic search was performed on isolated CHD and BW in PubMed, Embase, Web of Science, COCHRANE Library and Emcare. SELECTION CRITERIA: Neonates with isolated CHD were included if a BW percentile, BW z-score or % small-or-gestational age (SGA) was reported. DATA COLLECTION AND ANALYSIS: BW z-score and percentage SGA were pooled with random-effect meta-analysis. Quality and risk of bias were assessed using the modified Newcastle Ottawa Scale. MAIN RESULTS: Twenty-three articles (27 893 cases) were included. BW z-scores were retrieved from 11 articles, resulting in a pooled z-score of -0.20 (95% CI -0.50 to 0.11). The overall pooled prevalence of SGA <10th percentile was 16.0% (95% CI 11.4-20.5; 14 studies). Subgroup analysis of major CHD showed similar results (BW z-score -0.23 and percentage SGA 16.2%). CONCLUSIONS: Overall BW in isolated CHD is within range of normality but impaired, with a 1.6-fold higher risk of SGA, irrespective of the type of CHD (major CHD vs all CHD combined). Our findings underline the association between CHD and BW. The use of BW z-scores provides insight into growth of all fetuses with CHD. TWEETABLE ABSTRACT: Infants with a congenital heart defect (CHD) have a lower birthweight z-score and a higher incidence of small-for-gestational age (<10th percentile). This was encountered both in the major CHD-group as well as in all-CHD combined group analysis. Future research on the association between birthweight and CHD should include all types of CHDs (including mild cardiac defects) and placental-related disease, such as pre-eclampsia. We advocate the use of international standardised fetal growth and birthweight charts in CHD research.

Subcortical segmentation of the fetal brain in 3D ultrasound using deep learning.

The quantification of subcortical volume development from 3D fetal ultrasound can provide important diagnostic information during pregnancy monitoring. However, manual segmentation of subcortical structures in ultrasound volumes is time-consuming and challenging due to low soft tissue contrast, speckle and shadowing artifacts. For this reason, we developed a convolutional neural network (CNN) for the automated segmentation of the choroid plexus (CP), lateral posterior ventricle horns (LPVH), cavum septum pellucidum et vergae (CSPV), and cerebellum (CB) from 3D ultrasound. As ground-truth labels are scarce and expensive to obtain, we applied few-shot learning, in which only a small number of manual annotations (n = 9) are used to train a CNN. We compared training a CNN with only a few individually annotated volumes versus many weakly labelled volumes obtained from atlas-based segmentations. This showed that segmentation performance close to intra-observer variability can be obtained with only a handful of manual annotations. Finally, the trained models were applied to a large number (n = 278) of ultrasound image volumes of a diverse, healthy population, obtaining novel US-specific growth curves of the respective structures during the second trimester of gestation.

Placenta morphology and biomarkers in pregnancies with congenital heart disease - A systematic review.

Impaired placentation is an important contributing factor to intra-uterine growth restriction and pre-eclampsia in fetuses with congenital heart defects (CHD). These pregnancy complications occur more frequently in pregnancies with fetal CHD. One of the most important factors influencing the life of children with CHD is neurodevelopmental delay, which seems to start already in utero. Delayed neurodevelopment in utero may be correlated or even (partly) explained by impaired placentation in CHD cases. This systematic review provides an overview of published literature on placental development in pregnancies with fetal CHD. A systematic search was performed and the Newcastle-Ottawa scale was used to access data quality. Primary outcomes were placenta size and weight, vascular and villous architecture, immunohistochemistry, angiogenic biomarkers and/or placental gene expression. A total of 1161 articles were reviewed and 21 studies were included. Studies including CHD with a genetic disorder or syndrome and/or multiple pregnancies were excluded. Lower placental weight and elevated rates of abnormal umbilical cord insertions were found in CHD. Cases with CHD more frequently showed microscopic placental abnormalities (i.e. abnormal villous maturation and increased maternal vascular malperfusion lesions), reduced levels of angiogenic biomarkers and increased levels of anti-angiogenic biomarkers in maternal serum and umbilical cord blood. Altered gene expression involved in placental development and fetal growth were found in maternal serum and CHD placentas. In conclusion, abnormal placentation is found in CHD. More extensive studies are needed to elucidate the contribution of impaired placentation to delayed neurodevelopment in CHD cases.

Learning to map 2D ultrasound images into 3D space with minimal human annotation.

In fetal neurosonography, aligning two-dimensional (2D) ultrasound scans to their corresponding plane in the three-dimensional (3D) space remains a challenging task. In this paper, we propose a convolutional neural network that predicts the position of 2D ultrasound fetal brain scans in 3D atlas space. Instead of purely supervised learning that requires heavy annotations for each 2D scan, we train the model by sampling 2D slices from 3D fetal brain volumes, and target the model to predict the inverse of the sampling process, resembling the idea of self-supervised learning. We propose a model that takes a set of images as input, and learns to compare them in pairs. The pairwise comparison is weighted by the attention module based on its contribution to the prediction, which is learnt implicitly during training. The feature representation for each image is thus computed by incorporating the relative position information to all the other images in the set, and is later used for the final prediction. We benchmark our model on 2D slices sampled from 3D fetal brain volumes at 18-22 weeks' gestational age. Using three evaluation metrics, namely, Euclidean distance, plane angles and normalized cross correlation, which account for both the geometric and appearance discrepancy between the ground-truth and prediction, in all these metrics, our model outperforms a baseline model by as much as 23%, when the number of input images increases. We further demonstrate that our model generalizes to (i) real 2D standard transthalamic plane images, achieving comparable performance as human annotations, as well as (ii) video sequences of 2D freehand fetal brain scans.

Using three-dimensional ultrasound in predicting complex gastroschisis: A longitudinal, prospective, multicenter cohort study.

OBJECTIVE: To determine whether complex gastroschisis (ie, intestinal atresia, perforation, necrosis, or volvulus) can prenatally be distinguished from simple gastroschisis by fetal stomach volume and stomach-bladder distance, using three-dimensional (3D) ultrasound. METHODS: This multicenter prospective cohort study was conducted in the Netherlands between 2010 and 2015. Of seven university medical centers, we included the four centers that performed longitudinal 3D ultrasound measurements at a regular basis. We calculated stomach volumes (n = 223) using Sonography-based Automated Volume Count. The shortest stomach-bladder distance (n = 241) was determined using multiplanar visualization of the volume datasets. We used linear mixed modelling to evaluate the effect of gestational age and type of gastroschisis (simple or complex) on fetal stomach volume and stomach-bladder distance. RESULTS: We included 79 affected fetuses. Sixty-six (84%) had been assessed with 3D ultrasound at least once; 64 of these 66 were liveborn, nine (14%) had complex gastroschisis. With advancing gestational age, stomach volume significantly increased, and stomach-bladder distance decreased (both P < .001). The developmental changes did not differ significantly between fetuses with simple and complex gastroschisis, neither for fetal stomach volume (P = .85), nor for stomach bladder distance (P = .78). CONCLUSION: Fetal stomach volume and stomach-bladder distance, measured during pregnancy using 3D ultrasonography, do not predict complex gastroschisis.