Intra- and interobserver agreement for fetal cerebral measurements in 3D-ultrasonography

Automated Segmentation of Fetal Intracranial Volume in Three-Dimensional Ultrasound Using Deep Learning: Identifying Sex Differences in Prenatal Brain Development

The human brain undergoes major developmental changes during pregnancy. Three-dimensional (3D) ultrasound images allow for the opportunity to investigate typical prenatal brain development on a large scale. Transabdominal ultrasound can be challenging due to the small fetal brain and its movement, as well as multiple sweeps that may not yield high-quality images, especially when brain structures are unclear. By applying the latest developments in artificial intelligence for automated image processing allowing automated training of brain anatomy in these images retrieving reliable quantitative brain measurements becomes possible at a large scale. Here, we developed a convolutional neural network (CNN) model for automated segmentation of fetal intracranial volume (ICV) from 3D ultrasound. We applied the trained model in a large longitudinal population sample from the YOUth Baby and Child cohort measured at 20- and 30-week of gestational age to investigate biological sex differences in fetal ICV as a proof-of-principle and validation for our automated method (N = 2235 individuals with 43492 ultrasounds). A total of 168 annotated, randomly selected, good quality 3D ultrasound whole-brain images were included to train a 3D CNN for automated fetal ICV segmentation. A data augmentation strategy provided physical variation to train the network. K-fold cross-validation and Bayesian optimization were used for network selection and the ensemble-based system combined multiple networks to form the final ensemble network. The final ensemble network produced consistent and high-quality segmentations of ICV (Dice Similarity Coefficient (DSC) > 0.93, Hausdorff Distance (HD): HDvoxel < 4.6 voxels, and HDphysical < 1.4 mm). In addition, we developed an automated quality control procedure to include the ultrasound scans that successfully predicted ICV from all 43492 3D ultrasounds available in all individuals, no longer requiring manual selection of the best scan for analysis. Our trained model automatically retrieved ultrasounds with brain data and estimated ICV and ICV growth in 7672 (18%) of ultrasounds in 1762 participants that passed the automatic quality control procedure. Boys had significantly larger ICV at 20-weeks (81.7 ± 0.4 mL vs. 80.8 ± 0.5 mL; B = 2.86; p = 5.7e-14) and 30-weeks (257.0 ± 0.9 mL vs. 245.1 ± 0.9 mL; B = 12.35; p = 8.2e-27) of pregnancy, and more pronounced ICV growth than girls (delta growth 0.12 mL/day; p = 1.8e-5). Our automated artificial intelligence approach provides an opportunity to investigate fetal brain development on a much larger scale and to answer fundamental questions related to prenatal brain development.

A Clinical Approach to Semiautomated Three-Dimensional Fetal Brain Biometry-Comparing the Strengths and Weaknesses of Two Diagnostic Tools: 5DCNS+TM and SonoCNSTM

(1) Objective: We aimed to evaluate the accuracy and efficacy of AI-assisted biometric measurements of the fetal central nervous system (CNS) by comparing two semiautomatic postprocessing tools. We further aimed to discuss the additional value of semiautomatically generated sagittal and coronal planes of the CNS. (2) Methods: Three-dimensional (3D) volumes were analyzed with two semiautomatic software tools, 5DCNS+™ and SonoCNS™. The application of 5DCNS+™ results in nine planes (axial, coronal and sagittal) displayed in a single template; SonoCNS™ depicts three axial cutting sections. The tools were compared regarding automatic biometric measurement accuracy. (3) Results: A total of 129 fetuses were included for final analysis. Our data indicate that, in terms of the biometric quantification of head circumference (HC), biparietal diameter (BPD), transcerebellar diameter (TCD) and cisterna magna (CM), the accuracy of SonoCNS™ was higher with respect to the manual measurement of an experienced examiner compared to 5DCNS+™, whereas it was the other way around regarding the diameter of the posterior horn of the lateral ventricle (Vp). The inclusion of four orthogonal coronal views in 5DCNS+™ gives valuable information regarding spatial arrangements, particularly of midline structures. (4) Conclusions: Both tools were able to ease assessment of the intracranial anatomy, highlighting the additional value of automated algorithms in clinical use. SonoCNS™ showed a superior accuracy of plane reconstruction and biometry, but volume reconstruction using 5DCNS+™ provided more detailed information, which is needed for an entire neurosonogram as suggested by international guidelines.

Automatic measurements of fetal intracranial volume from 3D ultrasound scans

Three-dimensional fetal ultrasound is commonly used to study the volumetric development of brain structures. To date, only a limited number of automatic procedures for delineating the intracranial volume exist. Hence, intracranial volume measurements from three-dimensional ultrasound images are predominantly performed manually. Here, we present and validate an automated tool to extract the intracranial volume from three-dimensional fetal ultrasound scans. The procedure is based on the registration of a brain model to a subject brain. The intracranial volume of the subject is measured by applying the inverse of the final transformation to an intracranial mask of the brain model. The automatic measurements showed a high correlation with manual delineation of the same subjects at two gestational ages, namely, around 20 and 30 weeks (linear fitting R2(20 weeks) = 0.88, R2(30 weeks) = 0.77; Intraclass Correlation Coefficients: 20 weeks=0.94, 30 weeks = 0.84). Overall, the automatic intracranial volumes were larger than the manually delineated ones (84 ± 16 vs. 76 ± 15 cm3; and 274 ± 35 vs. 237 ± 28 cm3), probably due to differences in cerebellum delineation. Notably, the automated measurements reproduced both the non-linear pattern of fetal brain growth and the increased inter-subject variability for older fetuses. By contrast, there was some disagreement between the manual and automatic delineation concerning the size of sexual dimorphism differences. The method presented here provides a relatively efficient way to delineate volumes of fetal brain structures like the intracranial volume automatically. It can be used as a research tool to investigate these structures in large cohorts, which will ultimately aid in understanding fetal structural human brain development.

Three-dimensional ultrasound VOCAL combined with contrast-enhanced ultrasound: an alternative to contrast-enhanced magnetic resonance imaging for evaluating ablation of benign uterine lesions

OBJECTIVE

This study explores the feasibility and value of three-dimensional ultrasound virtual organ computer-aided analysis (3D-VOCAL) combined with contrast-enhanced ultrasound (CEUS) for measuring the non-perfused volume (NPV) after microwave ablation (MWA) of benign uterine lesions.

METHODS

Fifty-six patients with uterine myoma (UM) and adenomyosis (AM) treated with MWA were enrolled. NPV measurements were obtained postoperatively using two-dimensional CEUS (2D-CEUS), 3D-VOCAL combined with CEUS and three-dimensional contrast-enhanced magnetic resonance imaging (3D-CEMRI). Bland-Altman analysis and intraclass correlation coefficient (ICC) values were used to analyze the agreement of NPV measurements obtained via 2D-CEUS and the combined method with 3D-CEMRI. The inter- and intra-observer agreements of the NPV values obtained with all three methods were also analyzed.

RESULTS

Considering 3D-CEMRI as the standard, 3D-VOCAL showed greater agreement than 2D-CEUS and higher ICCs (ICC, 0.999 vs. 0.891) than 2D-CEUS for different lesion types and sizes of non-perfusion areas (p < 0.001 for all comparisons). NPV measurements obtained via 2 D-CEUS and 3 D-CEMRI differed significantly for AM and non-perfusion areas with maximum diameter ≥5 cm (p < 0.05) and showed no significant differences (p > 0.05) for UM and non-perfusion areas with maximum diameter <5 cm. The NPV measurements obtained via 3D-VOCAL and 3D-CEMRI did not differ significantly (p > 0.05). The intra- and inter-observer agreements of 3D-VOCAL measurements were better than those of 2D-CEUS and slightly lower than those of 3D-CEMRI.

CONCLUSIONS

3D-VOCAL combined with CEUS provides accurate estimates of NPV after MWA of benign uterine lesions, and offers a reliable, simple and efficient alternative to CEMRI.

Association between maternal oxygenation and brain growth in fetuses with left-sided cardiac obstructive lesions

OBJECTIVE

Impaired brain growth has been observed in fetuses with left-sided obstructive lesions (LSOL). Maternal oxygenation (MO) can alter fetal cerebral oxygenation and vascular reactivity. Our aim was to observe whether brain growth improves during MO in fetuses with LSOL.

METHODS

Forty-six fetuses with LSOL and 23 control fetuses were enrolled in this prospective longitudinal study. Fetuses with LSOL were subgrouped into those with MO (LSOL-MO, n = 23) and those without MO (LSOL-nMO, n = 23). Fetal head circumference (HC) and total intracranial volume (TIV) were evaluated serially at 4-week intervals. Brain biometry and growth were analyzed using linear mixed models adjusted for gestational age and sex. Spearman's correlation coefficients were calculated to identify baseline characteristics predictive of brain growth in the LSOL-MO group.

RESULTS

Duration of MO therapy had significant interaction effects on cerebral biometry in fetuses with LSOL. TIV increased more rapidly after 8 weeks of oxygen exposure and HC was larger after 16 weeks of oxygen exposure in the LSOL-MO group compared with the LSOL-nMO group (P &lt; 0.001). The change in TIV at the final time- point relative to the initial timepoint in the LSOL-MO group correlated negatively with the baseline pulsatility index of the middle cerebral artery (r = -0.58, P = 0.003) and baseline myocardial performance index of the left ventricle (r = -0.68, P &lt; 0.001).

CONCLUSIONS

TIV and HC increased faster in fetuses with LSOL which had MO compared with those that did not. Lower cerebral vascular resistance and preserved left heart function at baseline may predict greater cerebral biometric growth during MO. Additional research, including larger serial studies, is needed to confirm these preliminary findings and evaluate the clinical application of MO in this population. © 2022 International Society of Ultrasound in Obstetrics and Gynecology.

The YOUth study: Rationale, design, and study procedures

Behavioral development in children shows large inter-individual variation, and is driven by the interplay between biological, psychological, and environmental processes. However, there is still little insight into how these processes interact. The YOUth cohort specifically focuses on two core characteristics of behavioral development: social competence and self-regulation. Social competence refers to the ability to engage in meaningful interactions with others, whereas self-regulation is the ability to control one's emotions, behavior, and impulses, to balance between reactivity and control of the reaction, and to adjust to the prevailing environment. YOUth is an accelerated population-based longitudinal cohort study with repeated measurements, centering on two groups: YOUth Baby & Child and YOUth Child & Adolescent. YOUth Baby & Child aims to include 3,000 pregnant women, their partners and children, wheras YOUth Child & Adolescent aims to include 2,000 children aged between 8 and 10 years old and their parents. All participants will be followed for at least 6 years, and potentially longer. In this paper we describe in detail the design of this study, the population included, the determinants, intermediate neurocognitive measures and outcomes included in the study. Furthermore, we describe in detail the procedures of inclusion, informed consent, and study participation.

Intra- and interobserver agreement for fetal cerebral measurements in 3D-ultrasonography

The aim of this study is to evaluate intra- and interobserver agreement for measurement of intracranial, cerebellar, and thalamic volume with the Virtual Organ Computer-aided AnaLysis (VOCAL) technique in three-dimensional ultrasound images, in comparison to two-dimensional measurements of these brain structures. Three-dimensional ultrasound images of the brains of 80 fetuses at 20-24 weeks' gestational age were obtained from YOUth, a Dutch prospective cohort study. Two observers performed offline measurement of the occipitofrontal diameter, intracranial volume, transcerebellar diameter, cerebellar volume, and thalamic width, area, and volume, independently. VOCAL was used for calculation of the volumes. The two-way random, single measures intraclass correlation coefficient (ICC) was used for analysis of agreement and Bland-Altman plots were configured. Intra- and interobserver agreement was almost perfect for occipitofrontal diameter (intra ICC 0.88, 95% CI 0.82-0.92; inter ICC 0.91, 95% CI 0.85-0.94), intracranial volume (intra ICC 0.96, 95% CI 0.91-0.98; inter ICC 0.97, 95% CI 0.96-0.98) and transcerebellar diameter (intra ICC 0.91, 95% CI 0.86-0.94; inter ICC 0.86, 95% CI 0.78-0.910). For cerebellar volume, the intraobserver agreement was almost perfect (0.85, 95% CI 0.76-0.90), whereas the interobserver agreement was substantial (0.75, 95% CI 0.44-0.88). Agreement was only moderate for thalamic measurements. Bland-Altman plots for the volume measurements are normally distributed with acceptable mean differences and 95% limits of agreement. The intra- and interobserver agreement of the measurement of intracranial and cerebellar volume with VOCAL was almost perfect. These measurements are therefore reliable, and can be used to investigate fetal brain development. Thalamic measurements are not reliable enough.