Fetal Brain Imaging Research Articles

Transformers in the Womb: Swin-UNETR Takes on Fetal Brain Imaging.

Transformers in the Womb: Swin-UNETR Takes on Fetal Brain Imaging.

Ex vivo magnetic resonance imaging of the human fetal brain.

The fetal brain undergoes a dynamic process of development during gestation, marked by well-orchestrated events such as neuronal proliferation, migration, axonal outgrowth, and dendritic arborization, mainly elucidated through histological studies. Ex vivo magnetic resonance imaging (MRI) has emerged as a useful tool for 3D visualization of the developing fetal brain, serving as a complementary tool to traditional histology. In this review, we summarized the commonly employed ex vivo MRI techniques and their advances in fetal brain imaging, as well as a standard protocol for postmortem fetal brain specimen collection and fixation. We then provided an overview of ex vivo MRI-based studies on the fetal brain. According to our review, ex vivo T1- or T2-weighted structural MRI has contributed to the characterization of the anatomy of transient neuronal proliferative zones, the basal ganglia, and the cortex. Diffusion MRI related techniques, such as diffusion tensor imaging and tractography, have helped to investigate the microstructural patterns of fetal brain tissue, as well as the early emergence and development of neuronal migration pathways and white matter bundles. Ex vivo MRI findings have shown strong histological correlations, supporting the potential of MRI in evaluating the developmental events in the fetal brain. Postmortem MRI examinations have also demonstrated comparable, and in certain cases, superior performance to traditional autopsy in revealing fetal brain abnormalities. In conclusion, ex vivo fetal brain MRI is an invaluable tool that provides unique insights into the early stages of brain development.

Prediction of fetal brain gestational age using multihead attention with Xception

Accurate gestational age (GA) prediction is crucial for monitoring fetal development and ensuring optimal prenatal care. Traditional methods often face challenges in terms of precision and prediction efficiency. In this context, leveraging modern deep learning (DL) techniques is a promising solution. This paper introduces a novel DL approach for GA prediction using fetal brain images obtained via magnetic resonance imaging (MRI), which combines the strength of the Xception pretrained model with a multihead attention (MHA) mechanism. The proposed model was trained on a diverse dataset comprising 52,900 fetal brain images from 741 patients. The images encompass a GA ranging from 19 to 39 weeks. These pretrained models served as feature extraction components during the training process. The extracted features were subsequently used as the inputs of different configurable MHAs, which produced GA predictions in days. The proposed model achieved promising results with 8 attention heads, 32 dimensionality of the key space and 32 dimensionality of the value space, with an R-squared (R2) value of 96.5 %, a mean absolute error (MAE) of 3.80 days, and a Pearson correlation coefficient (PCC) of 98.50 % for the test set. Additionally, the 5-fold cross-validation results reinforce the model's reliability, with an average R2 of 95.94 %, an MAE of 3.61 days, and a PCC of 98.02 %. The proposed model excels in different anatomical views, notably the axial and sagittal views. A comparative analysis of multiple planes and a single plane highlights the effectiveness of the proposed model against other state-of-the-art (SOTA) models reported in the literature. The proposed model could help clinicians accurately predict GA.

High-resolution anatomical imaging of the fetal brain with a reduced field of view using outer volume suppression.

To achieve high-resolution fetal brain anatomical imaging without introducing image artifacts by reducing the FOV, and to demonstrate improved image quality compared to conventional full-FOV fetal brain imaging. Reduced FOV was achieved by applying outer volume suppression (OVS) pulses immediately prior to standard single-shot fast spin echo (SSFSE) imaging. In the OVS preparation, a saturation RF pulse followed by a gradient spoiler was repeated three times with optimized flip-angle weightings and a variable spoiler scheme to enhance signal suppression. Simulations and phantom and in-vivo experiments were performed to evaluate OVS performance. In-vivo high-resolution SSFSE images acquired using the proposed approach were compared with conventional and high-resolution SSFSE images with a full FOV, using image quality scores assessed by neuroradiologists and calculated image metrics. Excellent signal suppression in the saturation bands was confirmed in phantom and in-vivo experiments. High-resolution SSFSE images with a reduced FOV acquired using OVS demonstrated the improved depiction of brain structures without significant motion and blurring artifacts. The proposed method showed the highest image quality scores in the criteria of sharpness, contrast, and artifact and was selected as the best method based on overall image quality. The calculated image sharpness and tissue contrast ratio were also the highest with the proposed method. High-resolution fetal brain anatomical images acquired using a reduced FOV with OVS demonstrated improved image quality both qualitatively and quantitatively, suggesting the potential for enhanced diagnostic accuracy in detecting fetal brain abnormalities in utero.

Contribution of fetal blood sampling to determining the prognosis of congenital cytomegalovirus infections: a case-cohort study in Switzerland

Cytomegalovirus is responsible for the most common congenital infection, affecting 0.5% to 1.0% of live births in Europe. Congenital cytomegalovirus infection can be diagnosed during pregnancy by viral DNA amplification in the amniotic fluid, but the prognosis of fetuses without severe brain abnormalities remains difficult to establish on the basis of prenatal imaging alone. To identify predictors of moderate to severe symptomatic cytomegalovirus infection among fetal blood parameters and to propose an algorithm on the basis of these parameters and on prenatal imaging that would provide the best positive and negative predictive values. Fetal blood sampling at 21-28 weeks gestation was performed in fetuses with congenital cytomegalovirus infection confirmed by amniocentesis after maternal infection in the first-trimester or periconceptional period. We compared the levels of hemoglobin, thrombocytes, γ-glutamyl transpeptidase, aspartate aminotransferase, alanine aminotransferase, β2-microglobulin, immunoglobulins G and M, and cytomegalovirus DNA viral loads in amniotic fluid and fetal blood between those with moderate to severe symptomatic infection and those with asymptomatic to mild infection (median follow-up of 36 months for live births). Among 58 fetuses included, 25 (43%) had a moderate to severe symptomatic infection: 16 with severe cerebral abnormalities, 5 with multiple signs or symptoms at birth, 2 with bilateral sensorineural hearing loss, and 2 with neurodevelopmental delay. The values of thrombocytes, aspartate aminotransferase, β2 microglobulin, Immunoglobulin M, and cytomegalovirus viral loads differed significantly between fetuses with moderate to severe symptomatic infection and those with asymptomatic to mild infection. The optimal strategy to predict moderate to severe symptomatic infection was to first perform fetal brain imaging, followed by fetal blood sampling with the following cutoffs: thrombocytes <120,000/mL, viremia ≥5 log10/mL, and β2 microglobulin ≥12 mg/L). This recursive algorithm had a negative predictive value of 100% for moderately to severely symptomatic infection. The combination of thrombocytes, β2-microglobulin, and cytomegalovirus viral load in fetal blood can be used for prognosis determination, particularly in cytomegalovirus-infected fetuses without severe brain abnormalities at the time of prenatal diagnosis. Future studies should evaluate whether these parameters remain useful in infected fetuses who have been treated with valacyclovir before fetal blood sampling.

Perforation of cavum septi pellucidi in open spina bifida and need for hydrocephalus treatment by 1 year of age.

In-utero repair of an open neural tube defect (ONTD) reduces the risk of developing severe hydrocephalus postnatally. Perforation of the cavum septi pellucidi (CSP) may reflect increased intraventricular pressure in the fetal brain. We sought to evaluate the association of perforated CSP visualized on fetal imaging before and/or after in-utero ONTD repair with the eventual need for hydrocephalus treatment by 1 year of age. This was a retrospective cohort study of consecutive patients who underwent laparotomy-assisted fetoscopic ONTD repair between 2014 and 2021 at a single center. Eligibility criteria for surgery were based on those of the Management of Myelomeningocele Study (MOMS), although a maternal prepregnancy body mass index of up to 40 kg/m2 was allowed. Fetal brain imaging was performed with ultrasound and magnetic resonance imaging (MRI) at referral and 6 weeks postoperatively. Stored ultrasound and MRI scans were reviewed retrospectively to assess CSP integrity. Medical records were reviewed to determine whether hydrocephalus treatment was needed within 1 year of age. Parametric and non-parametric tests were used as appropriate to compare outcomes between cases with perforated CSP and those with intact CSP as determined on ultrasound at referral. Logistic regression analysis was performed to assess the predictive performance of various imaging markers for the need for hydrocephalus treatment. A total of 110 patients were included. Perforated CSP was identified in 20.6% and 22.6% of cases on preoperative ultrasound and MRI, respectively, and in 26.6% and 24.2% on postoperative ultrasound and MRI, respectively. Ventricular size increased between referral and after surgery (median, 11.00 (range, 5.89-21.45) mm vs 16.00 (range, 7.00-43.5) mm; P < 0.01), as did the proportion of cases with severe ventriculomegaly (ventricular width ≥ 15 mm) (12.7% vs 57.8%; P < 0.01). Complete CSP evaluation was achieved on preoperative ultrasound in 107 cases, of which 22 had a perforated CSP and 85 had an intact CSP. The perforated-CSP group presented with larger ventricles (mean, 14.32 ± 3.45 mm vs 10.37 ± 2.37 mm; P < 0.01) and a higher rate of severe ventriculomegaly (40.9% vs 5.9%; P < 0.01) compared to those with an intact CSP. The same trends were observed at 6 weeks postoperatively for mean ventricular size (median, 21.0 (range, 13.0-43.5) mm vs 14.3 (range,7.0-29.0) mm; P < 0.01) and severe ventriculomegaly (95.0% vs 46.8%; P < 0.01). Cases with a perforated CSP at referral had a lower rate of hindbrain herniation (HBH) reversal postoperatively (65.0% vs 88.6%; P = 0.01) and were more likely to require treatment for hydrocephalus (89.5% vs 22.7%; P < 0.01). The strongest predictor of the need for hydrocephalus treatment within 1 year of age was lack of HBH reversal on MRI (odds ratio (OR), 36.20 (95% CI, 5.96-219.12); P < 0.01) followed by perforated CSP on ultrasound at referral (OR, 23.40 (95% CI, 5.42-100.98); P < 0.01) and by perforated CSP at 6-week postoperative ultrasound (OR, 19.48 (95% CI, 5.68-66.68); P < 0.01). The detection of a perforated CSP in fetuses with ONTD can reliably identify those cases at highest risk for needing hydrocephalus treatment by 1 year of age. Evaluation of this brain structure can improve counseling of families considering fetal surgery for ONTD, in order to set appropriate expectations about postnatal outcome. © 2023 International Society of Ultrasound in Obstetrics and Gynecology.

Fetal-BET: Brain Extraction Tool for Fetal MRI.

Goal: In this study, we address the critical challenge of fetal brain extraction from MRI sequences. Fetal MRI has played a crucial role in prenatal neurodevelopmental studies and in advancing our knowledge of fetal brain development in-utero. Fetal brain extraction is a necessary first step in most computational fetal brain MRI pipelines. However, it poses significant challenges due to 1) non-standard fetal head positioning, 2) fetal movements during examination, and 3) vastly heterogeneous appearance of the developing fetal brain and the neighboring fetal and maternal anatomy across gestation, and with various sequences and scanning conditions. Development of a machine learning method to effectively address this task requires a large and rich labeled dataset that has not been previously available. Currently, there is no method for accurate fetal brain extraction on various fetal MRI sequences. Methods: In this work, we first built a large annotated dataset of approximately 72,000 2D fetal brain MRI images. Our dataset covers the three common MRI sequences including T2-weighted, diffusion-weighted, and functional MRI acquired with different scanners. These data include images of normal and pathological brains. Using this dataset, we developed and validated deep learning methods, by exploiting the power of the U-Net style architectures, the attention mechanism, feature learning across multiple MRI modalities, and data augmentation for fast, accurate, and generalizable automatic fetal brain extraction. Results: Evaluations on independent test data, including data available from other centers, show that our method achieves accurate brain extraction on heterogeneous test data acquired with different scanners, on pathological brains, and at various gestational stages. Conclusions:By leveraging rich information from diverse multi-modality fetal MRI data, our proposed deep learning solution enables precise delineation of the fetal brain on various fetal MRI sequences. The robustness of our deep learning model underscores its potential utility for fetal brain imaging.

Automated deep bottleneck residual 82-layered architecture with Bayesian optimization for the classification of brain and common maternal fetal ultrasound planes.

Despite a worldwide decline in maternal mortality over the past two decades, a significant gap persists between low- and high-income countries, with 94% of maternal mortality concentrated in low and middle-income nations. Ultrasound serves as a prevalent diagnostic tool in prenatal care for monitoring fetal growth and development. Nevertheless, acquiring standard fetal ultrasound planes with accurate anatomical structures proves challenging and time-intensive, even for skilled sonographers. Therefore, for determining common maternal fetuses from ultrasound images, an automated computer-aided diagnostic (CAD) system is required. A new residual bottleneck mechanism-based deep learning architecture has been proposed that includes 82 layers deep. The proposed architecture has added three residual blocks, each including two highway paths and one skip connection. In addition, a convolutional layer has been added of size 3 × 3 before each residual block. In the training process, several hyper parameters have been initialized using Bayesian optimization (BO) rather than manual initialization. Deep features are extracted from the average pooling layer and performed the classification. In the classification process, an increase occurred in the computational time; therefore, we proposed an improved search-based moth flame optimization algorithm for optimal feature selection. The data is then classified using neural network classifiers based on the selected features. The experimental phase involved the analysis of ultrasound images, specifically focusing on fetal brain and common maternal fetal images. The proposed method achieved 78.5% and 79.4% accuracy for brain fetal planes and common maternal fetal planes. Comparison with several pre-trained neural nets and state-of-the-art (SOTA) optimization algorithms shows improved accuracy.

Estimate of fetal brain temperature using proton resonance frequency thermometry during 3 Tesla fetal magnetic resonance imaging.

T2-weighted Single Shot Fast Spin Echo (SSFSE) scans at 3 Tesla (3T) are increasingly used to image fetal pathology due to their excellent tissue contrast resolution and signal-to-noise ratio (SNR). Temperature changes that may occur in response to radio frequency (RF) pulses used for these sequences at 3T have not been studied in human fetal brains. To evaluate the safety of T2-weighted SSFSE for fetal brains at 3T, magnetic resonance (MR) thermometry was used to measure relative temperature changes in a typical clinical fetal brain MR exam. Relative temperature was estimated using sets of gradient recalled echo (GRE) images acquired before and after T2-weighted SSFSE images which lasted 27.47±8.19 minutes. Thirty-one fetuses with cardiac abnormalities, and 20 healthy controls were included in this study. Fetal brain temperature was estimated by proton resonance frequency (PRF) thermometry and compared to the estimated temperature in the gluteal muscle of the mother. Seven scans with excessive motion were excluded. Local outlier factor (LOF) was performed to remove 12 additional scans with spurious phase measurements due to motion degradation and potential field drift. Linear regression was performed to determine if temperature changes are dependent on the rate of energy deposition during the scan. For the 32 participants used in the analysis, 17 with cardiac abnormalities and 15 healthy controls, the average relative fetal temperate change was 0.19±0.73 ℃ higher than the mother, with no correlation between relative temperature change and the rate of images acquired during the scans (regression coefficient =-0.05, R-squared =0.05, P=0.22, F-statistic =1.60). The difference in the relative temperature changes between the fetal brain and mother's gluteal tissue in the healthy controls was on average 0.08 ℃ lower and found not to be statistically different (P=0.76) to the group with cardiac abnormalities. Our results indicate that the estimated relative temperature changes of the fetal brain compared to the mother's gluteal tissue from RF pulses during the course of the T2-weighted SSFSE fetal MR exam are minimal. The differences in acquired phase between these regions through the exam were found not to be statistically different. These findings support that fetal brain imaging at 3T is within FDA limits and safe.

Longitudinal MRI in the context of in utero surgery for open spina bifida: A descriptive study.

Fetal surgery for open spina bifida (OSB) requires comprehensive preoperative assessment using imaging for appropriate patient selection and to evaluate postoperative efficacy and complications. We explored patient access and conduct of fetal magnetic resonance imaging (MRI) for prenatal assessment of OSB patients eligible for fetal surgery. We compared imaging acquisition and reporting to the International Society of Ultrasound in Obstetrics and Gynecology MRI performance guidelines. We surveyed access to fetal MRI for OSB in referring fetal medicine units (FMUs) in the UK and Ireland, and two NHS England specialist commissioned fetal surgery centers (FSCs) at University College London Hospital, and University Hospitals KU Leuven Belgium. To study MRI acquisition protocols, we retrospectively analyzed fetal MRI images before and after fetal surgery for OSB. MRI for fetal OSB was accessible with appropriate specialists available to supervise, perform, and report scans. The average time to arrange a fetal MRI appointment from request was 4 ± 3 days (range, 0-10), the average scan time available was 37 ± 16 min (range, 20-80 min), with 15 ± 11 min (range, 0-30 min) extra time to repeat sequences as required. Specific MRI acquisition protocols, and MRI reporting templates were available in only 32% and 18% of units, respectively. Satisfactory T2-weighted (T2W) brain imaging acquired in three orthogonal planes was achieved preoperatively in all centers, and 6 weeks postoperatively in 96% of FSCs and 78% of referring FMUs. However, for T2W spine image acquisition referring FMUs were less able to provide three orthogonal planes presurgery (98% FSC vs. 50% FMU, p < 0.001), and 6 weeks post-surgery (100% FSC vs. 48% FMU, p < 0.001). Other standard imaging recommendations such as T1-weighted (T1W), gradient echo (GE) or echoplanar fetal brain and spine imaging in one or two orthogonal planes were more likely available in FSCs compared to FMUs pre- and post-surgery (p < 0.001). There was timely access to supervised MRI for OSB fetal surgery assessment. However, the provision of images of the fetal brain and spine in sufficient orthogonal planes, which are required for determining eligibility and to determine the reversal of hindbrain herniation after fetal surgery, were less frequently acquired. Our evidence suggests the need for specific guidance in relation to fetal MRI for OSB. We propose an example guidance for MRI acquisition and reporting.

Stochastic-offset-enhanced restricted slice excitation and 180° refocusing designs with spatially non-linear ΔB0 shim array fields.

Developing a general framework with a novel stochastic offset strategy for the design of optimized RF pulses and time-varying spatially non-linear ΔB0 shim array fields for restricted slice excitation and refocusing with refined magnetization profiles within the intervals of the fixed voxels. Our framework uses the decomposition property of the Bloch equations to enable joint design of RF-pulses and shim array fields for restricted slice excitation and refocusing with auto-differentiation optimization. Bloch simulations are performed independently on orthogonal basis vectors, Mx, My, and Mz, which enables designs for arbitrary initial magnetizations. Requirements for refocusing pulse designs are derived from the extended phase graph formalism obviating time-consuming sub-voxel isochromatic simulations to model the effects of crusher gradients. To refine resultant slice-profiles because of voxelwise optimization functions, we propose an algorithm that stochastically offsets spatial points at which loss is computed during optimization. We first applied our proposed design framework to standard slice-selective excitation and refocusing pulses in the absence of non-linear ΔB0 shim array fields and compared them against pulses designed with Shinnar-Le Roux algorithm. Next, we demonstrated our technique in a simulated setup of fetal brain imaging in pregnancy for restricted-slice excitation and refocusing of the fetal brain. Our proposed framework for optimizing RF pulse and time-varying spatially non-linear ΔB0 shim array fields achieve high fidelity restricted-slice excitation and refocusing for fetal MRI, which could enable zoomed fast-spin-echo-MRI and other applications.

Automatic Localization of the Pons and Vermis on Fetal Brain MR Imaging Using a U-Net Deep Learning Model.

An MRI of the fetus can enhance the identification of perinatal developmental disorders, which improves the accuracy of ultrasound. Manual MRI measurements require training, time, and intra-variability concerns. Pediatric neuroradiologists are also in short supply. Our purpose was developing a deep learning model and pipeline for automatically identifying anatomic landmarks on the pons and vermis in fetal brain MR imaging and suggesting suitable images for measuring the pons and vermis. We retrospectively used 55 pregnant patients who underwent fetal brain MR imaging with a HASTE protocol. Pediatric neuroradiologists selected them for landmark annotation on sagittal single-shot T2-weighted images, and the clinically reliable method was used as the criterion standard for the measurement of the pons and vermis. A U-Net-based deep learning model was developed to automatically identify fetal brain anatomic landmarks, including the 2 anterior-posterior landmarks of the pons and 2 anterior-posterior and 2 superior-inferior landmarks of the vermis. Four-fold cross-validation was performed to test the accuracy of the model using randomly divided and sorted gestational age-divided data sets. A confidence score of model prediction was generated for each testing case. Overall, 85% of the testing results showed a ≥90% confidence, with a mean error of <2.22 mm, providing overall better estimation results with fewer errors and higher confidence scores. The anterior and posterior pons and anterior vermis showed better estimation (which means fewer errors in landmark localization) and accuracy and a higher confidence level than other landmarks. We also developed a graphic user interface for clinical use. This deep learning-facilitated pipeline practically shortens the time spent on selecting good-quality fetal brain images and performing anatomic measurements for radiologists.

Effect of magnet strength on fetal brain biometry - a single-center retrospective MRI-based cohort study.

Abnormal fetal brain measurements might affect clinical management and parental counseling. The effect of between-field-strength differences was not evaluated in quantitative fetal brain imaging until now. Our study aimed to compare fetal brain biometry measurements in 3.0T with 1.5T scanners. A retrospective cohort of 1150 low-risk fetuses scanned between 2012 and 2021, with apparently normal brain anatomy, were retrospectively evaluated for biometric measurements. The cohort included 1.5T (442 fetuses) and 3.0T scans (708 fetuses) of populations with comparable characteristics in the same tertiary medical center. Manually measured biometry included bi-parietal, fronto-occipital and trans-cerebellar diameters, length of the corpus-callosum, vermis height, and width. Measurements were then converted to centiles based on previously reported biometric reference charts. The 1.5T centiles were compared with the 3.0T centiles. No significant differences between centiles of bi-parietal diameter, trans-cerebellar diameter, or length of the corpus callosum between 1.5T and 3.0T scanners were found. Small absolute differences were found in the vermis height, with higher centiles in the 3.0T, compared to the 1.5T scanner (54.6th-centile, vs. 39.0th-centile, p < 0.001); less significant differences were found in vermis width centiles (46.9th-centile vs. 37.5th-centile, p = 0.03). Fronto-occipital diameter was higher in 1.5T than in the 3.0T scanner (66.0th-centile vs. 61.8th-centile, p = 0.02). The increasing use of 3.0T MRI for fetal imaging poses a potential bias when using 1.5T-based charts. We elucidate those biometric measurements are comparable, with relatively small between-field-strength differences, when using manual biometric measurements. Small inter-magnet differences can be related to higher spatial resolution with 3T scanners and may be substantial when evaluating small brain structures, such as the vermis.

Fetal brain imaging provides valuable information in cCMV infected infants

Background Congenital CMV infection (cCMV) is the most common congenital infection with 10–15% of cases developing symptomatic disease. Early antiviral treatment is of essence when symptomatic disease is suspected. Recently, the use of neonatal imaging has been implicated as a prognostic tool for long term sequalae among asymptomatic newborns at high risk. Even though neonatal MRI is commonly used in neonatal symptomatic cCMV disease, it is less often used in asymptomatic newborns, mainly due to cost, access and difficulty to perform. We have therefore developed an interest in assessing the use of fetal imaging as an alternative. Our primary aim was to compare the fetal and neonatal MRIs in a small cohort 10 asymptomatic neonates with congenital CMV infection. Methods We performed a single-center retrospective cohort study (case-series) on a convenience sample of children born from January 2014 to March 2021 with confirmed congenital CMV infection who had undergone both fetal and neonatal MRIs. We created a checklist of relevant cerebral abnormalities and asked 4 blinded radiologists to assess the MRIs (2 for each, namely fetal and neonatal) and then compared the findings between the fetal and neonatal imaging as well as the concordance in reporting of abnormalities within each category. Findings Overall concordance between prenatal and postnatal scans was high (70%). When comparing the two blinded reports for each MRI, we found high levels of concordance: 90% concordance for fetal MRIs and 100% for neonatal MRIs. The most common abnormalities identified in both fetal and neonatal scans were “abnormal white matter hyperintensity” and “subependymal cysts.” Interpretation Even though this is a small descriptive study, it indicates that fetal MRI could potentially provide us with similar information as neonatal imaging. This study could form the basis for subsequent larger future studies.

Application of Automatic Segmentation on Super-Resolution Reconstruction MR Images of the Abnormal Fetal Brain.

Fetal brain MR imaging is clinically used to characterize fetal brain abnormalities. Recently, algorithms have been proposed to reconstruct high-resolution 3D fetal brain volumes from 2D slices. By means of these reconstructions, convolutional neural networks have been developed for automatic image segmentation to avoid labor-intensive manual annotations, usually trained on data of normal fetal brains. Herein, we tested the performance of an algorithm specifically developed for segmentation of abnormal fetal brains. This was a single-center retrospective study on MR images of 16 fetuses with severe CNS anomalies (gestation, 21-39 weeks). T2-weighted 2D slices were converted to 3D volumes using a super-resolution reconstruction algorithm. The acquired volumetric data were then processed by a novel convolutional neural network to perform segmentations of white matter and the ventricular system and cerebellum. These were compared with manual segmentation using the Dice coefficient, Hausdorff distance (95th percentile), and volume difference. Using interquartile ranges, we identified outliers of these metrics and further analyzed them in detail. The mean Dice coefficient was 96.2%, 93.7%, and 94.7% for white matter and the ventricular system and cerebellum, respectively. The Hausdorff distance was 1.1, 2.3, and 1.6 mm, respectively. The volume difference was 1.6, 1.4, and 0.3 mL, respectively. Of the 126 measurements, there were 16 outliers among 5 fetuses, discussed on a case-by-case basis. Our novel segmentation algorithm obtained excellent results on MR images of fetuses with severe brain abnormalities. Analysis of the outliers shows the need to include pathologies underrepresented in the current data set. Quality control to prevent occasional errors is still needed.

Retracted: An Automated Deep Learning Model for the Cerebellum Segmentation from Fetal Brain Images.

[This retracts the article DOI: 10.1155/2022/8342767.].

Fetal heart rate evolution and brain imaging findings in preterm infants with severe cerebral palsy

Cerebral palsy is more common among preterm infants than among full-term infants. Although there is still no clear evidence that fetal heart rate monitoring effectively reduces cerebral palsy incidence, it is helpful to estimate the timing of brain injury leading to cerebral palsy and the causal relationship with delivery based on the fetal heart rate evolution patterns. Understanding the relationship between the timing and the type of brain injury can help to identify preventive measures in obstetrical care. This study aimed to examine the relationship between the timing of insults and the type of brain injury in preterm infants with severe cerebral palsy. This longitudinal study was based on a nationwide database for cerebral palsy. The data of infants with severe cerebral palsy (equivalent to levels 3-5 of the Gross Motor Function Classification System-Expanded and Revised), born between 2009 and 2014 at 28 to 33 weeks of gestation, were included. The intrapartum fetal heart rate evolution patterns were evaluated by 3 obstetricians blinded to clinical information other than gestational age at birth, and these were categorized after agreement by at least 2 of the 3 reviewers into (1) continuous bradycardia, (2) persistently nonreassuring (prenatal onset), (3) reassuring-prolonged deceleration, (4) Hon's pattern (intrapartum onset), (5) persistently reassuring (pre- or postnatal onset), and (6) unclassified. Infant brain magnetic resonance imaging findings at term-equivalent age were assessed by a pediatric neurologist blinded to the background details, except for gestational age at birth and corrected age at image acquisition, and these were categorized as (1) basal ganglia-thalamus, (2) white matter, (3) watershed cortex or subcortex, (4) stroke, (5) normal, and (6) unclassified based on the predominant site involved. The risk factors for the basal ganglia-thalamus group were compared with those of the combined white matter and watershed injuries group. Among 1593 infants with severe cerebral palsy, 231 were born at 28 to 33 weeks of gestation, and 140 met the eligibility criteria. Fetal heart rate evolution patterns were categorized as bradycardia (17% [24]); persistently nonreassuring (40% [56]); reassuring-prolonged deceleration (7% [10]); reassuring-Hon (6% [8]); persistently reassuring (7% [10]); and unclassified (23% [32]). Cerebral palsy was presumed to have an antenatal onset in 57% of infants and to have been caused by intrapartum insult in 13% of infants. Magnetic resonance imaging showed that 34% (n=48) of infants developed basal ganglia-thalamus-dominant brain injury. Of the remaining 92 infants, 43% (60) showed white matter injuries, 1% (1) showed watershed injuries, 4% (5) showed stroke, 1% (1) had normal findings, and 18% (25) had unclassified findings. Infants with continuous bradycardia (adjusted odds ratio, 1033.06; 95% confidence interval, 15.49-68,879.92) and persistently nonreassuring fetal heart rate patterns (61.20; 2.09-1793.12) had a significantly increased risk for basal ganglia-thalamus injury. Severe cerebral palsy was presumed to have an antenatal onset in 57% of infants and to have been caused by intrapartum insult in only 13% of infants born at 28 to 33 weeks of gestation. Although the white matter-watershed injury was predominant in the study populations, severe acute hypoxia-ischemia may be an important prenatal etiology of severe cerebral palsy in preterm infants.

Diffusion tensor imaging of fetal brain: principles, potential and limitations of promising technique.

Human brain development is a complex process that begins in the third week of gestation. During early development, the fetal brain undergoes dynamic morphological changes. These changes result from events such as neurogenesis, neuronal migration, synapse formation, axonal growth and myelination. Disruption of any of these processes is thought to be responsible for a wide array of different pathologies. Recent advances in magnetic resonance imaging, especially diffusion-weighted imaging and diffusion tensor imaging (DTI), have enabled characterization and evaluation of brain development in utero. In this review, aimed at practitioners involved in fetal medicine and high-risk pregnancies, we provide a comprehensive overview of fetal DTI studies focusing on characterization of early normal brain development as well as evaluation of brain pathology in utero. We also discuss the reliability and limitations of fetal brain DTI. © 2022 International Society of Ultrasound in Obstetrics and Gynecology.

Fetal and neonatal brain injury in twins complicated by twin anemia polycythemia sequence.

To determine the rate of fetal and neonatal brain lesions and define risk factors for such lesions in pregnancies complicated by Twin Anemia Polycythemia Sequence (TAPS). A retrospective cohort study of monochorionic twin pregnancies which were diagnosed with TAPS in a single tertiary medical center between 2013 and 2021. Pregnancies were followed with fetal brain neurosonogram every 2weeks and fetal brain MRI (magnetic resonance imaging) was performed when indicated at 28-32weeks of gestation; post-natal brain imaging included neonatal brain ultrasound. Pregnancies with pre- and post-natal brain lesions were compared to those without such findings. Overall, 23 monochorionic diamniotic pregnancies were diagnosed with TAPS over the study period resulting in perinatal survival of 91.3% (42/46). In 6/23 (26%) pregnancies and 7/46 (15.2%) fetuses pre- or post-natal brain lesions were detected, of whom five were the polycythemic twins and two were the anemic twins. Brain findings included intra-cerebral hemorrhage and ischemic lesions and were diagnosed prenatally in 6/7 (85.7%) cases. No risk factors for severe brain lesions were identified. TAPS may place the fetuses and neonates at increased risk for cerebral injuries. Incorporation of fetal brain imaging protocols may enhance precise prenatal diagnosis and allow for accurate parental counseling and post-natal care.

An Automated Deep Learning Model for the Cerebellum Segmentation from Fetal Brain Images.

Cerebellum measures taken from routinely obtained ultrasound (US) images have been frequently employed to determine gestational age and identify developing central nervous system's anatomical abnormalities. Standardized cerebellar assessments from large-scale clinical datasets are required to investigate correlations between the growing cerebellum and postnatal neurodevelopmental results. These studies could uncover structural abnormalities that could be employed as indicators to forecast neurodevelopmental and growth consequences. To achieve this, higher-throughput, precise, and impartial measures must be used to replace the existing human, semiautomatic, and advanced algorithms, which seem to be time-consuming and inaccurate. In this article, we presented an innovative deep learning (DL) technique for automatic fetal cerebellum segmentation from 2-dimensional (2D) US brain images. We present ReU-Net, a semantic segmentation network tailored to the anatomy of the fetal cerebellum. Moreover, we use U-Net as a foundation models with the incorporation of residual blocks and Wiener filter over the last 2 layers to segregate the cerebellum (c) from the noisy US data. 590 images for training and 150 images for testing were taken; also, we employed a 5-fold cross-assessment method. Our ReU-Net scored 91%, 92%, 25.42, 98%, 92%, and 94% for Dice Score Coefficient (DSC), F1-score, Hausdorff Distance (HD), accuracy, recall, and precision, correspondingly. The suggested method outperforms the other U-Net predicated techniques by a quantitatively significant margin (p 0.001). Our presented approach can be used to allow high bandwidth imaging techniques in medical study fetal US images as well as biometric evaluation on a broader scale in fetal US images.