Role of four-dimensional ultrasound (spatiotemporal image correlation and sonography-based automated volume count) in prenatal assessment of atrial morphology in cardiosplenic syndromes

Advances in the Application of Artificial Intelligence in Fetal Echocardiography

Congenital heart disease is a severe health risk for newborns. Early detection of abnormalities in fetal cardiac structure and function during pregnancy can help patients seek timely diagnostic and therapeutic advice, and early intervention planning can significantly improve fetal survival rates. Echocardiography is one of the most accessible and widely used diagnostic tools in the diagnosis of fetal congenital heart disease. However, traditional fetal echocardiography has limitations due to fetal, maternal, and ultrasound equipment factors and is highly dependent on the skill level of the operator. Artificial intelligence (AI) technology, with its rapid development utilizing advanced computer algorithms, has great potential to empower sonographers in time-saving and accurate diagnosis and to bridge the skill gap in different regions. In recent years, AI-assisted fetal echocardiography has been successfully applied to a wide range of ultrasound diagnoses. This review systematically reviews the applications of AI in the field of fetal echocardiography over the years in terms of image processing, biometrics, and disease diagnosis and provides an outlook for future research.

Fetal intelligent navigation echocardiography (FINE) has superior performance compared to manual navigation of the fetal heart by non-expert sonologists

OBJECTIVES

Manual and intelligent navigation (i.e. fetal intelligent navigation echocardiography or FINE) by the operator are two methods to obtain standard fetal cardiac views from spatiotemporal image correlation (STIC) volumes. The objective was to compare the performance between manual and intelligent navigation (FINE) of the fetal heart by non-expert sonologists.

METHODS

In this prospective observational study, ten sonologists underwent formal training on both navigational methods. Subsequently, they were tested on their ability to obtain nine cardiac views from five STIC volumes of normal fetal hearts (19-28 gestational weeks) using such methods. The following parameters were determined for both methods: (1) success rate of obtaining nine cardiac views; (2) mean time to obtain nine cardiac views per sonologist; and (3) maximum number of cardiac views successfully obtained for each STIC volume.

RESULTS

All fetal cardiac images obtained from 100 STIC volumes (50 for each navigational method) were reviewed by an expert in fetal echocardiography. Compared to manual navigation, FINE was associated with a significantly: (1) higher success rate of obtaining eight (excluding the abdomen view) appropriate cardiac views (92-100% vs. 56-88%; all p<0.05); (2) shorter mean time (minute:seconds) to obtain nine cardiac views (2:11 ± 0:37 vs. 15:49 ± 7:44; p<0.0001); and (3) higher success rate of obtaining all nine cardiac views for a given STIC volume (86 vs. 14%; p<0.001).

CONCLUSIONS

When performed by non-expert sonologists, intelligent navigation (FINE) had a superior performance compared to manual navigation of the normal fetal heart. Specifically, FINE obtained appropriate fetal cardiac views in 92-100% of cases.

Characterization of phenotypic spectrum of fetal heterotaxy syndrome by combining ultrasound and magnetic resonance imaging

OBJECTIVE

Heterotaxy or isomerism of the atrial appendages is a congenital disorder with variable presentation, associated with both cardiac and non-cardiac anomalies, which may have a serious impact on fetal outcome. The aim of this exploratory study was to assess the value of fetal magnetic resonance imaging (MRI), as a complementary tool to ultrasound, for describing the morphological spectrum encountered in heterotaxy.

METHODS

This retrospective study included 27 fetuses that underwent fetal MRI following prenatal suspicion of heterotaxy on ultrasound from 1998 to 2019 in a tertiary referral center. Heterotaxy was classified as left atrial isomerism (LAI) or right atrial isomerism (RAI) based on fetal echocardiography (FE) examination. In addition to routine prenatal ultrasound, fetal MRI was offered routinely to enhance the diagnosis of non-cardiac anomalies, which might have been missed on ultrasound. Prenatal findings on ultrasound, FE and MRI were reviewed systematically and compared with those of postnatal imaging and autopsy reports.

RESULTS

Twenty-seven fetuses with heterotaxy and cardiovascular pathology, of which 19 (70%) had LAI and eight (30%) had RAI, were included. Seven (7/19 (37%)) fetuses with LAI had normal intracardiac anatomy, whereas all fetuses with RAI had a cardiac malformation. All 27 fetuses had non-cardiac anomalies on fetal MRI, including situs and splenic anomalies. In 12/19 (63%) fetuses with LAI, a specific abnormal configuration of the liver was observed on MRI. In three fetuses, fetal MRI revealed signs of total anomalous pulmonary venous connection obstruction. An abnormal bronchial tree pattern was suspected on prenatal MRI in 6/19 (32%) fetuses with LAI and 3/8 (38%) fetuses with RAI.

CONCLUSIONS

Visualization on MRI of non-cardiac anomalies in fetuses with suspected heterotaxy is feasible and can assist the complex diagnosis of this condition, despite its limitations. This modality potentially enables differentiation of less severe cases from more complex ones, which may have a poorer prognosis. Fetal MRI can assist in prenatal counseling and planning postnatal management. © 2021 The Authors. Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology.

4D fetal echocardiography-An update

With the introduction of the electronic 4-dimensional and spatial-temporal image Correlation (e-STIC), it is now possible to obtain large volume datasets of the fetal heart that are virtually free of artifact. This allows the examiner to use a number of imaging modalities when recording the volumes that include two-dimensional real time, power and color Doppler, and B-flow images. Once the volumes are obtained, manipulation of the volume dataset allows the examiner to recreate views of the fetal heart that enable examination of cardiac anatomy. The value of this technology is that a volume of the fetal heart can be obtained, irrespective of the position of the fetus in utero, and manipulated to render images for interpretation and diagnosis. This article presents a summary of the various imaging techniques and provides clinical examples of its application used for prenatal diagnosis of congenital heart defects and abnormal cardiac function.

Outcome of prenatally diagnosed fetal heterotaxy: systematic review and meta-analysis

OBJECTIVES

The main aim of this systematic review was to evaluate the prevalence and type of associated anomalies in fetuses with heterotaxy diagnosed prenatally on ultrasound; the perinatal outcome of these fetuses was also studied.

METHODS

An electronic search of MEDLINE, EMBASE and CINAHL databases was performed. Only studies reporting the prenatal diagnosis of isomerism were included. Outcomes observed included associated cardiac and extracardiac anomalies, fetal arrhythmia, abnormal karyotype, type of surgical repair and perinatal mortality. The analysis was stratified according to the type of heterotaxy syndrome (left (LAI) or right (RAI) atrial isomerism). Meta-analyses of proportions were used to combine data. Quality assessment of the included studies was performed using the Newcastle-Ottawa Scale for cohort studies.

RESULTS

Sixteen studies (647 fetuses) were included in the analysis. Atrioventricular septal defect was the most common associated major cardiac anomaly found both in fetuses with LAI (pooled proportion (PP), 59.3% (95% CI, 44.0-73.7%)), with obstructive lesions of the right outflow tract occurring in 35.5% of these cases, and in fetuses with RAI (PP, 72.9% (95% CI, 60.4-83.7%)). Fetal arrhythmias occurred in 36.7% (95% CI, 26.9-47.2%) of cases with LAI and were mainly represented by complete atrioventricular block, while this finding was uncommon in cases with RAI (PP, 1.3% (95% CI, 0.2-3.2%)). Abnormal stomach and liver position were found, respectively, in 59.4% (95% CI, 38.1-79.0%) and 32.5% (95% CI, 11.9-57.6%) of cases with LAI, and in 54.5% (95% CI, 38.5-70.1%) and 45.9% (95% CI, 11.3-83.0%) of cases with RAI, while intestinal malrotation was detected in 14.2% (95% CI, 2.5-33.1%) of LAI and 27.1% (95% CI, 7.9-52.0%) of RAI cases. Hydrops developed in 11.8% (95% CI, 2.9-25.6%) of fetuses diagnosed prenatally with LAI. Biventricular repair was accomplished in 78.2% (95% CI, 64.3-89.4%) of cases with LAI, while univentricular repair or palliation was needed in 17.0% (95% CI, 9.7-25.9%); death during or after surgery occurred in 26.8% (95% CI, 4.6-58.7%) of LAI cases. Most children with RAI had univentricular repair and 27.8% (95% CI, 15.5-42.1%) died during or after surgery.

CONCLUSIONS

Fetal heterotaxy is associated with a high prevalence of cardiac and extracardiac anomalies. Approximately one quarter of fetuses with heterotaxy died during or after surgery. Abnormal heart rhythm, especially heart block, is common in fetuses with LAI, while this finding is uncommon in RAI. Biventricular repair was common in LAI while univentricular repair was required in the majority of children affected by RAI. Copyright © 2017 ISUOG. Published by John Wiley & Sons Ltd.

Three- and four-dimensional ultrasound in fetal echocardiography: an up-to-date overview

Congenital heart diseases (CHD) are the most commonly overlooked lesions in prenatal screening programs. Real-time two-dimensional ultrasound (2DUS) is the conventionally used tool for fetal echocardiography. Although continuous improvements in the hardware and post-processing software have resulted in a good image quality even in late first trimester, 2DUS still has its limitations. Four-dimensional ultrasound with spatiotemporal image correlation (STIC) is an automated volume acquisition, recording a single three-dimensional (3D) volume throughout a complete cardiac cycle, which results in a four-dimensional (4D) volume. STIC has the potential to increase the detection rate of CHD. The aim of this study is to provide a practical overview of the possibilities and (dis)advantages of STIC. A review of literature and evaluation of the current status and clinical value of 3D/4D ultrasound in prenatal screening and diagnosis of congenital heart disease are presented.

Prospective evaluation of the fetal heart using Fetal Intelligent Navigation Echocardiography (FINE)

OBJECTIVE

To evaluate prospectively the performance of Fetal Intelligent Navigation Echocardiography (FINE) applied to spatiotemporal image correlation (STIC) volume datasets of the normal fetal heart.

METHODS

In all women between 19 and 30 weeks' gestation with a normal fetal heart, an attempt was made to acquire STIC volume datasets of the apical four-chamber view if the following criteria were met: (1) fetal spine located between 5- and 7-o'clock positions; (2) minimal or absent shadowing (including a clearly visible transverse aortic arch); (3) absence of fetal breathing, hiccups, or movement; and (4) adequate image quality. Each STIC volume successfully acquired was evaluated by STICLoop™ to determine its appropriateness before applying the FINE method. Visualization rates of fetal echocardiography views using diagnostic planes and/or Virtual Intelligent Sonographer Assistance (VIS-Assistance®) were calculated.

RESULTS

One or more STIC volumes (365 in total) were obtained successfully in 72.5% (150/207) of women undergoing ultrasound examination. Of the 365 volumes evaluated by STICLoop, 351 (96.2%) were considered to be appropriate. From the 351 STIC volumes, only one STIC volume per patient (n = 150) was analyzed using the FINE method, and consequently nine fetal echocardiography views were generated in 76-100% of cases using diagnostic planes only, in 98-100% of cases using VIS-Assistance only, and in 98-100% of cases when using a combination of diagnostic planes and/or VIS-Assistance.

CONCLUSIONS

In women between 19 and 30 weeks' gestation with a normal fetal heart undergoing prospective sonographic examination, STIC volumes can be obtained successfully in 72.5% of cases. The FINE method can be applied to generate nine standard fetal echocardiography views in 98-100% of these cases using a combination of diagnostic planes and/or VIS-Assistance. This suggests that FINE could be implemented in fetal cardiac screening programs. Published 2015. This article is a U.S. Government work and is in the public domain in the USA.

Viewing rate and reference ranges for papillary muscle areas of the fetal heart using four-dimensional ultrasound in the rendering mode

OBJECTIVES

The aims of this study were to assess the viewing rate and determine reference ranges ​​for the papillary muscle areas in the fetal atrio-ventricular valves using four-dimensional (4D) ultrasound and spatio-temporal image correlation (STIC) in the rendering mode.

METHODS

This cross-sectional prospective study included 310 4D STIC volume data sets of normal fetuses between 18 weeks 0 day and 34 weeks 0 day of gestation. The papillary muscles were antero-lateral (MPAL) and postero-medial (MPPM) to the mitral valve and antero-superior (MPAS), inferior (MPI) and septal (MPS) to the tricuspid valve. Polynomial regressions were built to determine the reference ranges, and adjustments were made using the determination coefficient (R(2) ). To assess inter-observer reproducibility, the intra-class correlation coefficient (ICC) was used.

RESULTS

Identification of the papillary muscles was possible in 89.3% patients. The best-fit regression equations between papillary muscle areas and gestational age were second degree. The inter-observer reproducibility was good (ICC: 0.98 to MPAS, 0.97 to MPI, 0.98 to MPS, 0.98 to MPAL and 0.97 to MPPM).

CONCLUSIONS

Identification of the papillary muscles of the fetal valves was possible in most of the 4D STIC volume data sets, enabling the determination of reference ranges using the rendering mode. The reference ranges ​​for the papillary muscle areas were determined.

Real-time three-dimensional foetal echocardiography using a new transabdominal xMATRIX array transducer

BACKGROUND

Foetal echocardiography has been used to diagnose congenital heart disease. However, conventional echocardiography can only display two-dimensional (2D) structural images of the intricate three-dimensional (3D) foetal heart.

AIM

The purpose of this study was to report the first use of a new transabdominal xMATRIX array transducer and to describe its ability to perform all 3D modalities: intelligent spatiotemporal image correlation (iSTIC) acquisition, xPlane imaging and 3D surface imaging.

METHODS

Eighty foetuses without congenital heart disease were included consecutively, with a gestational age between 20 and 37 weeks. 2D and 3D scans were performed with a transabdominal xMATRIX array transducer. Cardiac-STIC volume datasets were acquired and postprocessed with new automatic software (the 'Fetal Heart Navigator').

RESULTS

A total of 224 iSTIC acquisitions were performed (mean time for each, 2 seconds). Only 78 iSTIC acquisitions (35%) were able to detect the ductal arch automatically. 'Fetal Heart Navigator' feasibility varied according to foetal position, including the descending aorta. Live xPlane imaging had excellent feasibility regardless of foetal position; using rotation, lateral and vertical tilts, all cardiac structures were identified from a unique reference plane. Live 3D surface imaging had variable feasibility depending on the target structure. Only 10% of the volume dataset offered comprehensive imaging of intracardiac views.

CONCLUSION

The new xMATRIX transabdominal transducer allows a multimodality approach to the foetal heart. Further studies that include foetuses with cardiac malformations are required.

Fetal Intelligent Navigation Echocardiography (FINE): a novel method for rapid, simple, and automatic examination of the fetal heart

OBJECTIVE

To describe a novel method (Fetal Intelligent Navigation Echocardiography (FINE)) for visualization of standard fetal echocardiography views from volume datasets obtained with spatiotemporal image correlation (STIC) and application of 'intelligent navigation' technology.

METHODS

We developed a method to: 1) demonstrate nine cardiac diagnostic planes; and 2) spontaneously navigate the anatomy surrounding each of the nine cardiac diagnostic planes (Virtual Intelligent Sonographer Assistance (VIS-Assistance®)). The method consists of marking seven anatomical structures of the fetal heart. The following echocardiography views are then automatically generated: 1) four chamber; 2) five chamber; 3) left ventricular outflow tract; 4) short-axis view of great vessels/right ventricular outflow tract; 5) three vessels and trachea; 6) abdomen/stomach; 7) ductal arch; 8) aortic arch; and 9) superior and inferior vena cava. The FINE method was tested in a separate set of 50 STIC volumes of normal hearts (18.6-37.2 weeks of gestation), and visualization rates for fetal echocardiography views using diagnostic planes and/or VIS-Assistance® were calculated. To examine the feasibility of identifying abnormal cardiac anatomy, we tested the method in four cases with proven congenital heart defects (coarctation of aorta, tetralogy of Fallot, transposition of great vessels and pulmonary atresia with intact ventricular septum).

RESULTS

In normal cases, the FINE method was able to generate nine fetal echocardiography views using: 1) diagnostic planes in 78-100% of cases; 2) VIS-Assistance® in 98-100% of cases; and 3) a combination of diagnostic planes and/or VIS-Assistance® in 98-100% of cases. In all four abnormal cases, the FINE method demonstrated evidence of abnormal fetal cardiac anatomy.

CONCLUSIONS

The FINE method can be used to visualize nine standard fetal echocardiography views in normal hearts by applying 'intelligent navigation' technology to STIC volume datasets. This method can simplify examination of the fetal heart and reduce operator dependency. The observation of abnormal echocardiography views in the diagnostic planes and/or VIS-Assistance® should raise the index of suspicion for congenital heart disease.

Embryo volume measurement: an intraobserver, intermethod comparative study of semiautomated and manual three-dimensional ultrasound techniques

OBJECTIVE

To compare the reliability of our recently introduced technique for first-trimester embryo volume measurement, the 'semiautomated technique' using both Virtual Organ Computer-aided AnaLysis (VOCAL(™) ) and Sonography-based Automated Volume Count (SonoAVC) with a manual technique using VOCAL alone.

METHODS

Seventy-four subjects with viable, singleton first-trimester in-vitro fertilization (IVF) pregnancies were recruited. Each subject underwent transvaginal sonography, at which a three-dimensional ultrasound dataset of the entire gestational sac was acquired. Each embryo volume was measured by two techniques, each performed twice. In the semiautomated technique VOCAL was used to calculate the volume of the gestational and yolk sacs, and SonoAVC was used to quantify the fluid volume within the amniotic and extra-amniotic cavities. Embryo volume was calculated by subtracting the sum of yolk sac, amniotic and extra-amniotic fluid volumes from gestational sac volume. In the manual technique, VOCAL was used to delineate the entire embryo using 9° rotations. Reliability was assessed using limits of agreement and intraclass correlation coefficient.

RESULTS

Datasets were included from 52 eligible subjects. Median gestational age was 7 + 4 weeks; median crown-rump length (CRL) was 13 (range, 2-29) mm; and median embryo volume was 1.8 (range, 0.03-8.1) cm(3) using the semiautomated technique and 0.7 (range 0.009-3.6) cm(3) using the manual technique. There was a significant discrepancy in the volumes measured by the two different techniques. Assessment of the limits of agreement suggested that the semiautomated technique (-15.8% to 14.0% of the mean embryo volume) was more reliable than was the manual technique (-22.4% to 26.6%).

CONCLUSION

The semiautomated technique is more reliable than is the manual technique for embryo volume measurement. However, the discrepancy in measurements between the two methods raises concerns over the validity of the semiautomated technique that require further investigation.