Automatic segmentation of 15 critical anatomical labels and measurements of cardiac axis and cardiothoracic ratio in fetal four chambers using nnU-NetV2.

BACKGROUND: Accurate segmentation of critical anatomical structures in fetal four-chamber view images is essential for the early detection of congenital heart defects. Current prenatal screening methods rely on manual measurements, which are time-consuming and prone to inter-observer variability. This study develops an AI-based model using the state-of-the-art nnU-NetV2 architecture for automatic segmentation and measurement of key anatomical structures in fetal four-chamber view images. METHODS: A dataset, consisting of 1,083 high-quality fetal four-chamber view images, was annotated with 15 critical anatomical labels and divided into training/validation (867 images) and test (216 images) sets. An AI-based model using the nnU-NetV2 architecture was trained on the annotated images and evaluated using the mean Dice coefficient (mDice) and mean intersection over union (mIoU) metrics. The model's performance in automatically computing the cardiac axis (CAx) and cardiothoracic ratio (CTR) was compared with measurements from sonographers with varying levels of experience. RESULTS: The AI-based model achieved a mDice coefficient of 87.11% and an mIoU of 77.68% for the segmentation of critical anatomical structures. The model's automated CAx and CTR measurements showed strong agreement with those of experienced sonographers, with respective intraclass correlation coefficients (ICCs) of 0.83 and 0.81. Bland-Altman analysis further confirmed the high agreement between the model and experienced sonographers. CONCLUSION: We developed an AI-based model using the nnU-NetV2 architecture for accurate segmentation and automated measurement of critical anatomical structures in fetal four-chamber view images. Our model demonstrated high segmentation accuracy and strong agreement with experienced sonographers in computing clinically relevant parameters. This approach has the potential to improve the efficiency and reliability of prenatal cardiac screening, ultimately contributing to the early detection of congenital heart defects.

Measurement of fetal automated fractional shortening using two-dimensional tracking in multiple centers.

PURPOSE: To establish a normal reference range for automated fractional shortening (Auto FS) in normal singleton fetuses measured at multiple centers. METHODS: This study was conducted from May 2017 to March 2019. It was undertaken on normal singleton fetuses. First, a four-chamber view of the fetal heart was recorded in the B-mode. Then, the region of interest was set on the edge of the ventricular septum and on the edge of the ventricular muscle at a point one-third away from the atrioventricular valve and toward the cardiac apex. Tracking was automatically performed. Values measured in the right ventricle were defined as R-Auto FS, and in the left ventricle as L-Auto FS. Furthermore, combined-Auto FS was defined as the measurement across both ventricles. RESULTS: A total of 442 normal fetuses were assessed. R-Auto FS decreased significantly with gestational age, and L-Auto FS showed a similar tendency (Spearman's correlation analysis: rs = - 0.415 and rs = - 0.252, respectively). Combined-Auto FS showed a similar decline as the gestational age increased (rs = - 0.451). CONCLUSION: In this study, we succeeded in defining a reference Auto FS value not only at one institution but also multiple centers. This study suggests that Auto FS can be used clinically and effectively.

The three fetal shunts: A story of wrong eponyms.

The fetal circulatory system bypasses the lungs and liver with three shunts. The foramen ovale allows the transfer of the blood from the right to the left atrium, and the ductus arteriosus permits the transfer of the blood from the pulmonary artery to the aorta. The ductus venosus is the continuation of the umbilical vein, allowing a large part of the oxygenated blood from the placenta to join the supradiaphragmatic inferior vena cava, bypassing the fetal liver and directly connecting the right atrium. These structures are named after the physicians who are thought to have discovered them. The foramen ovale and the ductus arteriosus are called the "foramen Botalli" and the "ductus Botalli," after Leonardo Botallo (1530-c. 1587). The ductus venosus is styled "ductus Arantii" after Giulio Cesare Arantius (1530-1589). However, these eponyms have been incorrectly applied as these structures were, in fact, discovered by others earlier. Indeed, the foramen ovale and the ductus arteriosus were described by Galen of Pergamon centuries earlier (c. 129-210 AD). He understood that these structures were peculiar to the fetal heart and that they undergo closure after birth. The ductus venosus was first described by Andreas Vesalius (1514-1564) 3 years before Arantius. Therefore, the current anatomical nomenclature of the fetal cardiac shunts is historically inappropriate.

The construction and application of an ultrasound and anatomical cross-sectional database of structural malformations of the fetal heart.

OBJECTIVES: Establish a fetal heart anatomical cross-sectional database that correlates with screening transverse ultrasound images suggested by international professional organizations to detect congenital heart defects. METHODS: Fetuses with suspected congenital heart defects identified using the following cardiac image sequences obtained from transverse slices beginning from the upper abdomen and ending in the upper thorax were the subjects of this study: (1) four-chamber view, (2) left ventricular outflow tract view, (3) three-vessel right ventricular outflow tract view, and (4) the three-vessel tracheal view. A database of digital two-dimensional images of the transverse sweep was created for fetuses with confirmed congenital heart defects. In addition, using four-dimensional ultrasound spatial-temporal image correlation, selected transverse ultrasound images were acquired as part of the database. Ultrasound-detected congenital heart defects were confirmed postnatally from pathological specimens of the heart and lungs using a cross-sectional technique that mirrored the ultrasound images described above. When anatomical specimens were not available, prenatal ultrasound-detected congenital heart defects were confirmed using postnatal echocardiography and/or following surgery. RESULTS: The four screening views described in the Methods section identified 160 fetuses that comprised the database. Forty-five datasets consisted of both ultrasound and anatomical cross-sectional images. Thirteen percent (6/45) only had abnormalities of the four-chamber view (eg, endocardial cushion defects). Twenty-four percent (11/45) had abnormalities of the four-chamber view as well as right and left outflow tracts (eg, complex malformations). Of these, 10 of 11 had an abnormal tracheal view. Sixteen percent (7/45) had an abnormal four-chamber view and abnormal right outflow tract (eg, pulmonary stenosis). Thirty-three percent (15/45) had a normal four-chamber view but had abnormal right and left outflow tracts as well as an abnormal tracheal view (eg, tetralogy of Fallot, D-transposition of the great arteries). CONCLUSIONS: Combining both ultrasound and anatomical imaging may be of assistance in training imagers to recognize cardiovascular pathology when performing the screening examination of the fetal heart.

DW-Net: A cascaded convolutional neural network for apical four-chamber view segmentation in fetal echocardiography.

Fetal echocardiography (FE) is a widely used medical examination for early diagnosis of congenital heart disease (CHD). The apical four-chamber view (A4C) is an important view among early FE images. Accurate segmentation of crucial anatomical structures in the A4C view is a useful and important step for early diagnosis and timely treatment of CHDs. However, it is a challenging task due to several unfavorable factors: (a) artifacts and speckle noise produced by ultrasound imaging. (b) category confusion caused by the similarity of anatomical structures and variations of scanning angles. (c) missing boundaries. In this paper, we propose an end-to-end DW-Net for accurate segmentation of seven important anatomical structures in the A4C view. The network comprises two components: 1) a Dilated Convolutional Chain (DCC) for "gridding issue" reduction, multi-scale contextual information aggregation and accurate localization of cardiac chambers. 2) a W-Net for gaining more precise boundaries and yielding refined segmentation results. Extensive experiments of the proposed method on a dataset of 895 A4C views have demonstrated that DW-Net can achieve good segmentation results, including the Dice Similarity Coefficient (DSC) of 0.827, the Pixel Accuracy (PA) of 0.933, the AUC of 0.990 and it substantially outperformed some well-known segmentation methods. Our work was highly valued by experienced clinicians. The accurate and automatic segmentation of the A4C view using the proposed DW-Net can benefit further extractions of useful clinical indicators in early FE and improve the prenatal diagnostic accuracy and efficiency of CHDs.

Free-breathing fetal cardiac MRI with doppler ultrasound gating, compressed sensing, and motion compensation.

BACKGROUND: Fetal cardiovascular MRI complements ultrasound to assess fetal cardiovascular pathophysiology. PURPOSE: To develop a free-breathing method for retrospective fetal cine MRI using Doppler ultrasound (DUS) cardiac gating and tiny golden angle radial sampling (tyGRASP) for accelerated acquisition capable of detecting fetal movements for motion compensation. STUDY TYPE: Feasibility study. SUBJECTS: Nine volunteers (gestational week 34-40). Short-axis and four-chamber views were acquired during maternal free-breathing and breath-hold. FIELD STRENGTH/SEQUENCE: 1.5T cine balanced steady-state free precession. ASSESSMENT: A self-gated reconstruction method was improved for clinical application by using 1) retrospective DUS gating, and 2) motion detection and rejection/correction algorithms for compensating for fetal motion. The free-breathing reconstructions were qualitatively and quantitatively assessed, and DUS-gating was compared with self-gating in breath-hold reconstructions. A scoring of 1-4 for overall image quality, cardiac, and extracardiac diagnostic quality was used. STATISTICAL TESTS: Friedman's test was used to assess differences in qualitative scoring between observers. A Wilcoxon matched-pairs signed rank test was used to assess differences between breath-hold and free-breathing acquisitions and between observers' quantitative measurements. RESULTS: In all cases, 111 free-breathing and 145 breath-hold acquisitions, the automatically calculated DUS-based cardiac gating signal provided reconstructions of diagnostic quality (median score 4, range 1-4). Free-breathing did not affect the DUS-based cardiac gated retrospective radial reconstruction with respect to image or diagnostic quality (all P > 0.06). Motion detection with rejection/correction in k-space produced high-quality free-breathing DUS-based reconstructions [median 3, range (2-4)], whereas free-breathing self-gated methods failed in 80 out of 88 cases to produce a stable gating signal. DATA CONCLUSION: Free-breathing fetal cine cardiac MRI based on DUS gating and tyGRASP with motion compensation yields diagnostic images. This simplifies acquisition for the pregnant woman and thus could help increase fetal cardiac MRI acceptance in the clinic. LEVEL OF EVIDENCE: 2 Technical Efficacy Stage: 1 J. Magn. Reson. Imaging 2020;51:260-272.

The morphological investigations on the heart and some vessels of bovine foetus between the 15th and 25th weeks of gestation.

BACKGROUND: The aim of this study was to define the morphological and morphometric development of the foetus heart obtained from the domestic cattle in the gestation period of 15-25 weeks. MATERIALS AND METHODS: For this purpose, a total of 30 hearts belonging to cattle foetuses (15 males, 15 females) were used. The ages of foetuses were calculated according to the forehead-to-tail length and examined in three different groups. After dissection; biometric, macroanatomic, morphometric and histological findings were obtained from the foetal hearts according to the groups. In addition, mean values of the morphometric findings were determined. RESULTS: As a result of the study, it was found that with the advancing age the convexity of margo ventricularis dexter increased and margo ventricularis sinister transformed from a convex-concave shape to a flat shape. The heart-to-body weight ratio was determined as 0.08% for Group II female foetuses and 0.09% for all other groups. The heart heights for Groups I, II, and III females were identified as 26.21, 41.00, and 46.27 mm, respectively, and for the males 26.45, 34.89, and 47.15 mm, respectively. In the statistical analysis, it was determined that all the morphometric values measured from the heart correlated significantly with the forehead-to-tail length. CONCLUSIONS: The data obtained as a result of the study is thought to help understand the morphological and morphometrical development of the heart, pioneer the attempts to create a foetal cattle heart model, and thus help in the diagnosis of the foetal heart pathologies.acielecka.

Utility of Fetal Cardiac Axis and Cardiac Position Assessment in Predicting Neonatal Respiratory Morbidity in Fetal Congenital Lung Lesions.

OBJECTIVES: To assess the diagnostic performance of the fetal cardiac axis (CA) and/or cardiac position (CP) versus the congenital pulmonary malformation volume ratio (CVR) in predicting any and severe neonatal respiratory morbidity in fetal congenital lung lesions. METHODS: This work was an 11-year retrospective cohort study. The sensitivity, specificity, positive predictive value, and negative predictive value of CA and/or CP assessment in prediction of respiratory morbidity were calculated before 24 weeks' gestation and between 24 and 32 weeks and compared to CVR cutoffs obtained from the literature. RESULTS: Fifty-three patients were included. CA and/or CP abnormalities were present in 45% and 38% of patients before 24 weeks and between 24 and 32 weeks and were significantly more common in left- versus right-sided lesions (60% versus 17%; P = .003). The sensitivity, specificity, positive predictive value, and negative predictive value of an abnormal CA and/or CP for any and severe respiratory morbidity were 0.67, 0.61, 0.33, and 0.86 and 0.8, 0.58, 0.17, and 0.97 before 24 weeks and 0.75, 0.73, 0.45, and 0.91 and 0.8, 0.67, 0.20, and 0.97 between 24 and 32 weeks, respectively. An abnormal CA and/or CP had higher sensitivity for any respiratory morbidity compared to the CVR at 0.5 and 0.8 cutoffs both before 24 weeks and between 24 and 32 weeks (P < .05). CONCLUSIONS: An abnormal CA and/or CP before 24 weeks and between 24 and 32 weeks has higher sensitivity for the detection of any respiratory morbidity at birth compared to the CVR at both 0.5 and 0.8 cutoffs. A normal CA and CP have a high negative predictive value for excluding any respiratory morbidity at birth both before 24 weeks and between 24 and 32 weeks.

Comparison of 2D versus M-mode echocardiography for assessing fetal myocardial wall thickness.

OBJECTIVE: M-mode and 2D have been proposed for evaluating fetal myocardial thickness. However, studies comparing the performance of both modalities are lacking. We aimed to compare 2D versus M-mode reproducibility for assessing myocardial wall thicknesses. METHODS: A prospective study including 45 healthy fetuses from low-risk pregnancies evaluated between 18 and 41 weeks of gestation. Left and right ventricular free-wall and septal myocardial thicknesses were measured at end-diastole (ED) and end-systole (ES) in transverse 4-chamber view using 2D and M-mode. Intra- and interobserver reproducibility was evaluated by the concordance correlation coefficient (CCC). Both techniques were compared by t-test of the CCC. RESULTS: 2D and M-mode demonstrated excellent and similar intraobserver repeatability, with the best concordance in ES septal thickness (M-mode CCC 0.956 versus 2D-mode CCC 0.914). Interobserver reproducibility demonstrated also a high concordance, optimal in ES left ventricular free wall (M-mode 0.925 versus 2 D 0.855). Comparison of both techniques demonstrated a high concordance in all measurements, except for ED septal thickness with better reproducibility using M-mode (CCC 0.954 versus 0.847, p = .017). CONCLUSIONS: 2D and M-mode can be used in a reproducible manner for measuring fetal myocardial thickness, with a slightly better performance of M-mode for assessing ED septal wall thickness.

Observations of foetal heart veins draining directly into the left and right atria.

Evaluation of semiserial sections of 14 normal hearts from human foetuses of gestational age 25-33 weeks showed that all of these hearts contained thin veins draining directly into the atria (maximum, 10 veins per heart). Of the 75 veins in these 14 hearts, 55 emptied into the right atrium and 20 into the left atrium. These veins were not accompanied by nerves, in contrast to tributaries of the great cardiac vein, and were negative for both smooth muscle actin (SMA) and CD34. However, the epithelium and venous wall of the anterior cardiac vein, the thickest of the direct draining veins, were strongly positive for SMA and CD34, respectively. In general, developing fibres in the vascular wall were positive for CD34, while the endothelium of the arteries and veins was strongly positive for the present DAKO antibody of SMA. The small cardiac vein, a thin but permanent tributary of the terminal portion of the great cardiac vein, was also positive for SMA and CD34. A few S100 protein-positive nerves were observed along both the anterior and small cardiac veins, but no nerves accompanied the direct dra- inage veins. These findings suggested that the latter did not develop from the early epicardiac vascular plexus but from a gulfing of the intratrabecular space or sinus of the atria. However, the immunoreactivity of the anterior cardiac vein suggests that it originated from the vascular plexus, similar to tributaries of the great cardiac vein.

Z Score Reference Ranges of Fetal Cardiothoracic Diameter Ratio.

OBJECTIVE: This study aims to establish the z score reference ranges of cardiothoracic diameter ratio (CTR) of normal fetuses from 14 to 40 weeks' gestation. METHOD: A cross-sectional study was conducted on low-risk singleton pregnancies with healthy fetuses. The CTR was performed using simple 2-dimensional sonography on the typical fetal 4-chamber view. The reference ranges were constructed according to gestational age (GA), biparietal diameter (BPD), and head circumference (HC) as independent variables based on the best-fit models, both mean and standard deviation (SD). RESULTS: A total of 683 fetuses were measured for CTR. The best-fit equations for the mean and SD as a function of GA, BPD, and HC are as follows: (1) CTR = 0.365 + 0.004 × GA in weeks (SD = 0.031 + 0.001 × GA); (2) CTR = 0.373 + 0.014 × BPD in cm (SD = 0.034 + 0.004 × BPD); and (3) CTR = 0.373 + 0.004 × HC in cm (SD = 0.032 + 0.001 × HC). The CTR was slightly increased with advanced GA, BPD, and HC. CONCLUSION: Complete nomograms with z score reference ranges of CTR were established throughout pregnancy. These nomograms may be useful to detect cardiac abnormalities.

Evaluation of fetal heart geometry during pregnancy by three-dimensional ultrasound using the STIC rendering mode.

OBJECTIVE: To determine fetal heart geometry during pregnancy using three-dimensional (3D) ultrasound and the spatiotemporal image correlation (STIC) rendering mode. METHODS: This prospective, cross-sectional study evaluated 250 normal singleton pregnancies from 20 to 33 weeks and 6 days of gestation. STIC rendering was used to calculate the eight angles of the fetal heart: apex, base, mitral valve, tricuspid valve, left ventricle, right ventricle, left atrium, and right atrium angles. The concordance correlation coefficient (CCC) was used for intra- and inter-observer tests. RESULTS: The average ± SD maternal age was 31.7 ± 4.9 years, and the average gestational age was 26.3 ± 4.2 weeks. There was little variation in fetal heart angles using STIC rendering according to the gestational age, with determination coefficient (R2 ) values of 0.01 for the apex and mitral valve angles and <0.01 for the base, tricuspid valve, left ventricle, right ventricle, left atrium, and right atrium angles. Moderate/good intra- and inter-observer concordance was observed for the measurement of fetal heart angles using STIC rendering, and the obtained CCC varied from 0.74 to 0.93. CONCLUSION: The fetal heart geometry did not present significant variations during pregnancy using 3D ultrasound and the STIC rendering mode.

24-segment sphericity index: a new technique to evaluate fetal cardiac diastolic shape.

OBJECTIVE: Because of parallel circulation in the fetus and the differential effect that various disease states may have on the shape of the right and left ventricles, this study was conducted to evaluate the sphericity index (SI) of 24 transverse segments distributed from the base to the apex of each of the ventricular chambers. METHODS: Two hundred control fetuses were examined between 20 and 40 weeks of gestation. The displacement of the ventricular endocardium during the cardiac cycle was computed using offline speckle-tracking software. From the ASCII output of the analysis, we analyzed 24 end-diastolic transverse segments, distributed from the base to the apex of each ventricle, as well as the end-diastolic mid-basal-apical length. The SI was computed for each of the 24 segments by dividing the mid-basal-apical length by the transverse length for each segment. Regression analysis was performed against biometric measurements and gestational age according to last menstrual period and ultrasound. Eight fetuses, in which the four-chamber view appeared subjectively to demonstrate chamber disproportion, were evaluated as examples to demonstrate the utility of this technology. RESULTS: The SI for each segment was independent of gestational age and fetal biometric measurements. The SI of the right ventricle was significantly (P < 0.001) lower than that of the left ventricle for segments 1-18, suggesting that the right ventricle was more globular in shape than was the left ventricle at the base, mid and a portion of the apical segments of the chamber. Fetuses with various cardiac structural abnormalities and abnormal fetal growth had abnormal SI values that reflected either a more globular or a more flattened ventricular chamber. CONCLUSION: Determination of SI for each of 24 segments of the fetal right and left ventricles provides a comprehensive method to examine the shape of the ventricular chambers. Copyright © 2017 ISUOG. Published by John Wiley & Sons Ltd.

Correlation Between the Thymic-Thoracic Ratio and Cardiac Axis in Healthy Fetuses.

OBJECTIVES: The aim of this study was to assess the correlation between the thymic-thoracic ratio and cardiac axis in healthy fetuses. METHODS: The fetal thymic-thoracic ratio and cardiac axis were measured in 220 healthy fetuses. The normal ranges of the fetal thymic-thoracic ratio and cardiac axis for each gestational week were established. The Pearson correlation coefficient was calculated to analyze the relationship between the thymic-thoracic ratio and cardiac axis. RESULTS: The mean fetal thymic-thoracic ratio increased slightly throughout the pregnancy, from 0.33 at 17 weeks' gestation to 0.37 at 37 weeks. No statistically significant correlation was found between the thymic-thoracic ratio and cardiac axis in healthy fetuses. CONCLUSIONS: In healthy fetuses, the thymus occupies an increasingly larger portion of the upper mediastinum when the pregnancy advances. The cardiac axis is not affected by the thymic-thoracic ratio in healthy fetuses.

Common Findings in Late-Gestation Fetal Echocardiography.

OBJECTIVES: Fetal echocardiography provides detailed information about cardiac structure and function in utero. Limited information is available regarding normal findings late in pregnancy. We therefore sought to identify and describe common cardiac findings in late gestation. METHODS: Fetuses with structurally normal hearts were identified in mid gestation within a subset of pregnant women in a prospective study. The atrioventricular valves, right and left atria, aortic isthmus and ductus arteriosus dimensions and flow abnormalities, aneurysm of the septum primum, and presence and grade of tricuspid regurgitation were assessed throughout pregnancy. Linear and logistic regression analyses were used to characterize changes in quantitative and qualitative fetal echocardiographic parameters by gestational age (GA). RESULTS: Forty fetuses between 24 and 38 weeks' GA were studied. Each had a fetal echocardiographic study completed before and after 34 weeks' gestation, which were compared. Tricuspid-to-mitral valve and right-to-left atrium ratios increased with GA (P < .001). More frequently noted after 34 weeks were tapering of the ductus arteriosus (2.5% versus 32%), prominent aortic isthmus diastolic flow (5% versus 67%), prominent ductus arteriosus diastolic flow (2.5% versus 25%), trivial or mild tricuspid regurgitation (35% versus 80%), and aneurysms of the septum primum (37.5% versus 80%). These findings all increased linearly with GA (P < .001). CONCLUSIONS: Atrioventricular valve and right/left atrium disproportion, mild ductus arteriosus tapering, prominent aortic isthmus and ductus arteriosus diastolic flow, trivial or mild tricuspid regurgitation, and aneurysms of the septum primum are frequently identified after 34 weeks' GA. Their identification suggests that these fetal echocardiographic findings in isolation are likely normal and are results of the physiologic alterations that occur late in the third trimester.

Perinatal Changes in Fetal Ventricular Geometry, Myocardial Performance, and Cardiac Function in Normal Term Pregnancies.

BACKGROUND: The fetal heart at term is exposed to an increase in hemodynamic work as a consequence of fetal growth, increased circulating volume, and alteration in loading patterns due to maturational changes in fetoplacental circulation. The extent to which these cardiovascular changes influence human fetal and neonatal cardiac adaptation has not been fully elucidated. The aim of this study was to evaluate perinatal cardiovascular changes in ventricular geometry and myocardial performance in normal term fetuses. METHODS: Prospective study of 108 uncomplicated pregnancies delivering at term. M-mode, two-dimensional or B-mode, pulsed wave Doppler, pulsed wave tissue Doppler, and two-dimensional speckle-tracking imaging were performed a few days before and within 24 hours of birth. RESULTS: Analysis of paired fetal and neonatal echoes demonstrated significant perinatal changes (P < .0001 for all) in right ventricular (RV) and left ventricular (LV) geometry (RV/LV end-diastolic dimension ratio, 1.2 vs 0.8; RV sphericity index, 0.53 vs 0.40; LV sphericity index, 0.46 vs 0.49). There were corresponding significant (P < .001 for all) perinatal changes in global myocardial performance: LV myocardial performance index, 0.60 versus 0.47; RV myocardial performance index, 0.61 versus 0.42; systolic function: LV longitudinal systolic strain rate, -1.4/sec versus -1.0/sec; RV longitudinal systolic strain rate, -1.5/sec versus -1.0/sec; RV S', 5.3 cm/sec versus 6.5 cm/sec; and diastolic function: LV E'/A', 0.8 versus 1.1. CONCLUSIONS: The findings support the concept that the perinatal period is associated with major changes in fetal ventricular geometry and cardiac function in response to significant alterations in loading conditions. Improved knowledge of perinatal cardiac changes in normal fetuses could facilitate better understanding of cardiac adaptation in normal and pathological pregnancies.