Rendering in fetal cardiac scanning: the intracardiac septa and the coronal atrioventricular valve planes

Authors


Abstract

Objective

In this study we aimed to apply spatio-temporal image correlation (STIC) rendering to visualize the virtual planes of the interventricular and interatrial septa (IVS, IAS) as well as the atrioventricular (AV) annuli plane just distal to the semilunar valves (coronal atrioventricular (CAV) plane) in normal and pathological fetal hearts, to ascertain whether these planes add to fetal cardiac examination.

Methods

Unselected gravidae presenting for anatomy scan or patients referred for fetal echocardiography in the second and third trimesters of pregnancy with suspected or diagnosed cardiac malformation were scanned using the five planes technique with the STIC modality to obtain cardiac volume sets for each patient. Rendering capabilities were employed to obtain the ‘virtual planes’ to evaluate the IVS, IAS, AV annuli, and size and alignment of the great vessels.

Results

A total of 136 normal scans were performed to establish a learning curve for STIC acquisition and post-processing rendering and analysis. An additional 35 cases with cardiac anomalies were accrued. In 131/136 (96.3%) normal scans the IAS and IVS were visualized successfully, while in 127/136 (93.4%) normal fetuses the CAV plane was successfully visualized. In 13 anomalous cases the IVS plane improved ventricular septal defect (VSD) evaluation, and in four the IAS plane contributed to foramen ovale evaluation. The modality improved visualization of the septa and the assessment of the defects, as well as the foramen ovale flap and pattern of movement of the foramen ovale. In five cases the CAV plane improved evaluation of the alignment of the major vessels in relation to the AV annuli, and in three the evaluation of the semilunar valves, with or without malalignment of the great vessels.

Conclusions

Rendering STIC technology allows the visualization of virtual planes (IAS, IVS, AV annuli–CAV plane), which can clarify our understanding of anatomical defects and may improve communication with the management team and family. Copyright © 2006 ISUOG. Published by John Wiley & Sons, Ltd.

Introduction

Congenital heart defects (CHDs) are the most common congenital anatomic malformation1, 2. They affect approximately 0.6–5% of liveborn children2–5 and their prevalence in abortuses has been shown to be much higher6, 7. Despite great efforts and considerable progress in both the technology and technique of fetal echocardiographic scanning over the past two decades, the accuracy of prenatal diagnosis of CHD ranges from 31 to 96%2, 8–15, and many normal and anomalous fetal cardiac structures have not been fully delineated echocardiographically, such as the lateral view of the interatrial septum (IAS) and interventricular septum (IVS), the anteroposterior view of the atrioventricular (AV) annuli and the alignment of the great vessels. This stems from the technical difficulty of imaging these planes adequately for evaluation of these structures by two-dimensional ultrasound (2DUS) scanning.

Ventricular septal defects (VSDs) are among those most often missed at prenatal scanning3–5, 13. Though many smaller defects may close spontaneously, larger or multiple lesions can have significant impact on neonatal management16. The four-chamber view of the heart with color Doppler mapping alone is often not sufficient to detect VSDs, as the lesions can lie at any level of the three-dimensional septum, and blood flow across the lesion can be intermittent or absent because of fetal hemodynamics. Employing 2DUS in real time the operator cannot image the septum in its entirety. Using three-dimensional (3D) ultrasonographic spatio-temporal image correlation (STIC) acquisition with post-processing rendering capabilities we show here our ability to obtain ‘virtual planes’ of the entire ventricular septum, which is viewed as if from one side of the septum, looking towards the contralateral side.

The fetal atrial septum and foramen ovale have been studied and normal Doppler flow patterns described17. However, atrial septal defects (ASDs) are often missed in prenatal scanning because of confusion with the foramen ovale, and Doppler studies may be difficult owing to physiological blood flow across the foramen ovale. As with the IVS, 2DUS cannot image the complete IAS; we here demonstrate the virtual plane of the IAS obtained with STIC acquisition and post-processing application.

Maldevelopment of AV alignment and great vessels' relative sizes form a diverse group of cardiac anomalies affecting the AV alignments and connections, AV valves, and the arterioventricular alignments and connections. We demonstrate virtual planes of the AV septum—an anteroposterior plane of the AV annuli and the coronal atrioventricular (CAV) plane just distal to the great vessel valves. These virtual planes are not accessible by 2DUS; analysis of the acquired volume set allowed us to obtain these views to evaluate the AV alignments and connections and blood flow across the valves, as well as blood flow in the great vessels and their relative sizes.

By applying the recently developed STIC technology, through sweep acquisition of the volume of interest, and using multiplanar reconstruction to analyze the volume, the heart may be visualized not only in three orthogonal planes, but also in virtual planes not normally accessible with two-dimensional (2D) echocardiography.

In this study we aimed to apply STIC rendering to visualize the IVS and IAS as well as the AV annuli plane just distal to the semilunar valves (CAV plane) in normal and pathological fetal hearts, to ascertain whether these virtual planes add to the fetal cardiac examination.

Patients and Methods

During the study period July 2004 to July 2005, 136 unselected gravidae presenting for mid-trimester anatomy scan were examined with the five planes technique of fetal echocardiography18–21, followed by STIC acquisition22–31 of the same region of interest. These scans were analyzed to establish a learning curve for STIC acquisition and postprocessing manipulation of the volume set. In addition to the classic five planes, these volume sets were analyzed to obtain the IVS, IAS, and AV annuli–CAV planes.

The Necker and Mt Scopus centers perform fetal echocardiography in some 300 patients referred with suspected, and 100 patients referred with incidentally diagnosed, cases of congenital heart disease annually. Thirty-five patients presenting during the study period with suspected or diagnosed fetal cardiac anomaly who agreed to 3D examination were examined as above, and the volume sets analyzed to obtain and evaluate these planes as compared to the normal scans. All the STIC scans were performed by one examiner (S. Y.).

Volume acquisition and rendering

To acquire heart volumes we used a four-dimensional (4D) ultrasound system with STIC capability (Voluson 730 Expert, GE Medical Systems, Kretz, Austria) with a motorized 4–8-MHz curved-array transducer. Whenever possible the scan was performed with the fetus in a quiet state, without movement. Using transverse planes of the fetal upper abdomen and mediastinum in a continual sweep, automatic acquisition was performed lasting 10–15 s, with angles of acquisition between 15° and 35°. Acquisition was performed with and without color Doppler scanning.

The volume sets obtained were analyzed using the Voluson 4D-View postprocessing system. This is performed away from the patient on a dedicated PC. Having obtained the classic four-chamber view of the fetal heart in the upper left (A) panel of the multiplanar rendering screen, the operator adjusts the angle of the IVS and IAS to 90°, using Z-rotation. The operator then defines the rendering box as tightly as possible around the septa for their entire length. The side of the box showing a green line is the ‘active’ side: this will determine whether the septum is visualized from the right side of the heart toward the left, or vice-versa. Figure 1 shows acquisition of the normal IVS and IAS. Viewed from within the right ventricle the IVS will appear rougher while viewed from the left side the septum appears smoother (Figure 2).

Figure 1.

Ultrasound image of normal intracardiac septum. The rendering box is placed tightly around the septum in the four-chamber view acquired with spatio-temporal image correlation. (a) The side of the rendering box marked with a green line is ‘active’ and determines whether the septum is imaged from the left ventricle to the right or the reverse. (b) The rendered image of the interventricular septum (IVS) from within the left ventricle, with the fully opened foramen ovale. (c) The septum with the rendering box active from within the right ventricle. (d) The rendered image. Caret indicates the annulus in the area of the crux. FO, foramen ovale.

Figure 2.

(a) The interventricular septum imaged from within the right ventricle appears rough, showing the characteristic trabeculations. (b) The rendered image from within the left ventricle showing the relative smoothness of the ventricular wall as compared to (a) and the fully opened foramen ovale (FO). Insets show corresponding two-dimensional images.

To obtain the CAV plane, the operator once again begins from a good four-chamber view in panel A, as above. The rendering box is placed around the AV annuli. The view is then fine-tuned using X-axis rotation to give an anteroposterior plane through the heart (Figure 3). The rendering box is adjusted slightly coronally to observe blood flow in the great vessels.

Figure 3.

Ultrasound image in the normal coronal atrioventricular plane. The rendering box is placed tightly around the level of the atrioventricular (AV) valves and fine-tuned with the X-rotation option to image the great vessel valves as they begin to open and close. (a) Rendered image of the heart in mid-diastole: note the fully opened AV valves and closed aortic and pulmonary valves. (b) Rendered image in end-diastole: the tricuspid and mitral valves are closed, while the aortic and pulmonary valves are beginning to open. Note the aortic valve cusp just visible in the orifice. AO, aorta; PA, pulmonary artery; MV, mitral valve; TV, tricuspid valve.

Neonates were screened within 72 hours of delivery; suspected cases of congenital cardiac anomalies or murmurs were followed clinically for up to 30 days, including echocardiography when cardiac malformation or dysfunction was suspected. Neonatal echocardiography was performed with an HP-5500 machine (Hewlett-Packard, Palo-Alto, CA, USA), using an 8-MHz transducer.

Results

One hundred and thirty-six unselected gravidae and 35 cases with cardiac anomaly were examined during the study period. Among the former—gravidae from a general obstetric population presenting for routine mid-trimester anatomy scans at a mean gestational age of 22.5 (range, 21–26) weeks at our center—no cases of cardiac anomaly were diagnosed. Mean maternal age of this group was 26.5 (range, 19–46) years, and parity 1.7 (range, 0–8). In 131/136 scans (96.3%) the IVS and IAS were visualized successfully in postprocessing. In 127/136 cases (93.4%) the AV annuli–CAV plane was visualized successfully. Limitations to the examination were often the presence of excessive fetal activity or fetal breathing movements. In cases of fetal activity the operator would wait for a period of quiescence; with fetal breathing movements it was found that up to two such movements during acquisition did not have a significant adverse effect on scan quality. In failed cases STIC acquisition and subsequent visualization of the four-chamber view were suboptimal owing to maternal body habitus or excessive fetal movement of duration beyond reasonable extension of the examination. During analysis it was found that the CAV plane, as a ‘virtual’ plane through the fetal heart, was not uniformly visualized on a straight plane. Some fine-tuning with X- and Y-rotation was necessary; however, this did not adversely affect diagnostic acuity. Table 1 summarizes the prenatal and postnatal findings in the pathological cases, as well as the contribution of the described planes to the evaluation of these anomalies. In 13 cases evaluation of the IVS plane added to the visualization of VSD, including two cases where it excluded VSD. In four cases the IAS plane added to our evaluation of the foramen ovale, and in five cases the CAV plane improved visualization of alignment of the great vessels. In addition, application of these views facilitated evaluation of the semilunar valves and cardiac cushion in atrioventricular septal defects (AVSDs). Figures 4–7 show the IVS, IAS and CAV planes applied to the evaluation of cases of cardiac malformation.

Figure 4.

Rendered image of the interventricular septum showing ventricular septal defect (VSD) during diastole (a) and systole (b). Inset shows the corresponding two-dimensional image, in which the arrow indicates the VSD, and carets indicate the pulmonary veins. FO, foramen ovale.

Figure 5.

Rendered image of the interatrial septum showing restrictive foramen ovale (FO). This was the maximum opening of the FO observed in this fetus. The Rashkind procedure was performed 14 h after delivery; postnatal echocardiography diagnosed d-transposition of the great arteries with normal pulmonary valves. A small ventricular septal defect was confirmed during surgery.

Figure 6.

Ultrasound image in the coronal atrioventricular (CAV) plane showing pulmonary stenosis (PS). Note the retrograde flow in the critically stenotic main pulmonary artery. Right inset shows the normal CAV plane with color Doppler; left inset shows the two-dimensional gray-scale image with PS indicated by an arrow. AO, aorta; Lt, left; MV, mitral valve; PA, pulmonary artery; Rt, right; TV, tricuspid valve.

Figure 7.

(a) Ultrasound image in the coronal atrioventricular plane showing complete atrioventricular canal defect (AVC). (b) The defect in two dimensions. AO, aorta; AVSD, atrioventricular septal defect; PA, pulmonary artery.

Table 1. Summary of 35 anomalous cases evaluated with the interatrial septum (IAS), interventricular septum (IVS) and coronal atrioventricular (CAV) planes, and the benefit derived
CaseGA (weeks)Prenatal diagnosis by two-dimensional ultrasoundPrenatal diagnosis by three-dimensional ultrasound; benefit gainedNeonatal evaluation/ pathology exam
  • *

    Cases that benefited from additional evaluation with virtual planes. AO, aorta; ASD, atrial septal defect; AV, atrioventricular; AVSD, atrioventricular septal defect; CS, coronary sinus; FO, foramen ovale; GA, gestational age at diagnosis or referral; HLH, hypoplastic left heart; HRH, hypoplastic right heart; IUFD, intrauterine fetal death; PLSVC, persistent left superior vena cava; PM, post-mortem; PS, pulmonary stenosis; RV, right ventricle; TGA, transposition of great arteries; TOP, termination of pregnancy; TR, tricuspid regurgitation; TTTS, twin-to-twin transfusion syndrome; UVH, univentricular heart; VSD, ventricular septal defect.

 133 + 5d-TGA with VSDVisualization of VSD*d-TGA with VSD
 226d-TGA with VSDVisualization of VSD*d-TGA with VSD
 336d-TGAVisualization of IVS*d-TGA without VSD
 433d-TGAEvaluation of the FO*d-TGA with small VSD
 532d-TGARestrictive FO*d-TGA with restrictive FO, without VSD
 626 + 3d-TGA without VSDCAV plane; evaluation of IVS excluded VSD*d-TGA, without VSD
 722 + 3d-TGA with VSDEvaluation of VSD and FO*d-TGA with VSD
 821 + 1d-TGA with AV canalAlignment of great vessels*TOP; PM confirmed ultrasound findings
 923d-TGA, AVSDConfirmed two-dimensional ultrasoundConfirmed
1034 + 5Malposition of the great arteries with VSD and hypoplastic aortaVisualization of VSD*Two competent AV valves, large VSD, arterial malposition
1127Malposition of great arteries with VSD and hypoplastic aortaEvaluation of VSD, alignment of great vessels*Arterial malposition, large VSD, no coarctation
1227Pulmonary stenosis, apical VSDVisualization of VSD*Apical VSD, quasipulmonary atresia, right ventricular suprasystemic pressures
1339Agenesis of pulmonary valve with VSDNo significant compression of trachea in fetal lifeAgenesis of pulmonary valve with VSD; tracheal compression by endoscopy at 1 month of age
1429Coarctation of AO with VSDNo straddling of the AV valvesAortic hypoplasia with VSD
1525 + 3Aortic stenosis, small left ventricleEvaluation of AV valvesConfirmed postnatally
1638Maldeveloped left ventricle and vessels; PLSVC to CS; sus coarctation of AOThickened AO valve, detection of PLSVC to CS*Bicuspid AO valve with ASD; PLSVC to CS without coarctation of AO
1736 + 5Pulmonary atresia with intact septumNoninformative exam, excessive fetal movementPulmonary atresia with intact septum
1824UVH, pulmonary atresiaNo benefit: poor acquisitionTOP; UVH, pulmonary atresia
1935 + 6Severe pulmonary stenosis, tetralogy of Fallot, VSDAlignment of great vessels, evaluation of VSD*Severe PS, tetralogy of Fallot, VSD, trisomy 18, demise
2032 + 0Pulmonary stenosisVisualization of pulmonary valve*Pulmonary stenosis
2126 + 0Pulmonary atresia with VSD, maldeveloped right ventricleVisualization of VSD*TOP; PM confirmed ultrasound findings
2225 + 3Severe pulmonary stenosis, TTTS recipientVisualization of pulmonary valve*Pulmonary stenosis, moderate TR
2328 + 5VSD, small pulmonary artery and enlarged aortaEvaluation of VSD and great vesselsTOP; PM confirmed tetralogy of Fallot, trisomy 13
2421 + 0Tetralogy of FallotAlignment of great vessels, CAV plane and IVS*Tetralogy of Fallot confirmed postnatally
2535Complete AVSDCAV planeComplete AVSD
2616 + 5AVSDCAV planeTOP; PM refused
2727 + 4AVSDCAV planeAVSD confirmed postnatally, trisomy 21
2818 + 6HLHConfirmed two-dimensional ultrasoundTOP; PM confirmed ultrasound findings
2916 + 4HRH, tricuspid atresia in twin BConfirmed two-dimensional ultrasoundSelective termination
3021 + 3HLH with two VSDsVisualization of VSDs*TOP; PM confirmed ultrasound findings
3134Small RV, abnormal tricuspid valveNo benefit: poor acquisitionIsolated restrictive right ventricle without hemodynamic impairment
3232Asymmetry of ventricles, large FOBetter visualization of the FO*Normal at postpartum day 3
3314 + 6Heterotaxy (right-sided stomach), d-TGA, pulmonary atresia, AVSDAlignment of great vessels*TOP; PM refused
3429 + 3Truncus arteriosus, subtruncal VSD, intra-abdominal umbilical vein varixEvaluation of the septum and VSD*IUFD at 32 weeks: PM confirmed diagnosis, 22q11 microdeletion
3530Multiple hyperechogenic foci (> 15)Evaluation of valve and functionNormal heart

Neonatal screening of the control group revealed two small VSDs that were missed prenatally, as well as five cases of mild tricuspid regurgitation.

Discussion

CHDs are the most common congenital anomalies, and most cases occur in low-risk women1. Effective tools for prenatal detection of these anomalies are therefore essential. To be an effective tool fetal echocardiography must be reliable and clinically feasible, i.e. both accessible to operators and acceptable to patients. To this end, we and others have described the five planes of fetal echocardiography18–21, an effective method of streamlining fetal echocardiography scanning. For a technique to be effective, it is necessary that it be readily learnt by operators, and that the results help elucidate the nature of anomalies to interdisciplinary management teams, counselors and parents.

The goal of prenatal echocardiography is to visualize structural and functional anomalies of the fetal heart and great vessels to optimize diagnostic precision and to provide images that will aid both management teams and parents in understanding the nature of anomalies. The application of three-dimensional ultrasound (3DUS) rendering capabilities to fetal echocardiography improves the visualization of both normal and anomalous anatomy and function, and provides clearer, more understandable images for professionals and parents.

We show here that the intracardiac septum view, which provides a complete lateral image of the septum as recently described32, can be consistently visualized: in over 130 normal hearts it was successfully imaged in 96%, and images differentiate between the left and right sides of the septum. We also present the anteroposterior view of the AV annuli and great vessel valves, a plane not amenable to imaging in other modalities. This was also consistently visualized, and successfully obtained and evaluated in 93% of normal cases.

In 13 cases of anomalous fetal hearts, the intracardiac septum view improved evaluation of the VSD and in four the foramen ovale. In five anomalous cases, the CAV plane was important in the evaluation of the AV valves and alignment of the great vessels and in three the evaluation of the semilunar valves (Table 1, cases marked with an asterisk).

The views presented here of the IVS, IAS and the CAV plane just distal to the AV valves, extend our understanding and capabilities of arriving at optimal diagnostic accuracy, and of optimizing imaging for management teams and parental counseling. Visualization of the IVS increases diagnostic accuracy of the size and functionality of VSDs, while visualization of the IAS reveals the degree of motion or restriction of the foramen ovale flap. The ability to differentiate between the right and left faces of the IVS adds to our fuller evaluation of the defect, and may add to our understanding of the pathophysiology of these lesions.

The CAV plane provides images not formerly obtained of the relative positions of the great vessels, and the characteristics of anomalous connections of the vessels, chambers and AV valves. The contribution of the CAV plane seems to lie in evaluation of the relative positions of the great vessels, though it may not improve absolute rates of diagnosis of these lesions. It is of importance in cases of complex congenital heart defects to choose between different postnatal strategies such as biventricular repair and planned palliation with cavopulmonary shunts.

Among the most prevalent of CHDs are VSDs and ASDs, which represent some 42% and 8.6% of CHDs diagnosed in liveborn infants: these defects may be diagnosed in 0.25 and 0.05% of neonates, respectively4. Some studies report an incidence of up to 2–5% of examined neonates affected with VSD3. Many cases of small, nonfunctional VSD and ASD undoubtedly remain undiagnosed both prenatally and in the neonatal period; spontaneous closure is also not uncommon, and may occur in some 85% of cases3, 16. Septal defects are often seen with other cardiac or extracardiac defects, and must raise suspicion of concomitant anomaly when they are diagnosed.

The main obstacles to improved rates of diagnosis of ASD and VSD are confusion with the foramen ovale in the atrial septum, and minimal shunting of blood across the defects, owing to lower pressure gradients between the right and left fetal heart, as opposed to postpartum. For example, in 2DUS scanning it is difficult to image the three-dimensional IVS in all necessary planes to show the color Doppler jet across the defect to confirm VSD; color Doppler mapping of the classic four-chamber view plane may not disclose the defect if it occurs on another plane, or if the jet is intermittent. Further, the functional sequelae of ASDs and VSDs may not be apparent in utero.

However, it may be found that beyond the purely technical aspect, these planes may aid surgical teams, as was shown by Jouannic et al.33 in their study on the effect of prenatal diagnosis on neonatal clinical status, in cases of transposition of the great arteries. They found that a restrictive foramen ovale and/or ductus arteriosus may predict the need for emergency neonatal care for these fetuses. Precise delineation of anomalous anatomy of the AV valves and alignment of the great vessels may also find their place in the armamentarium of presurgical evaluation, as has been shown regarding the postnatal benefit of prenatal evaluation of other cardiac anomalies33–36. Similarly, improved visualization of the nature of changes in septal thickness and texture as well as the size of very large VSDs may assist parents in understanding the implications for an affected pregnancy and the likelihood of closure.

While this study encompasses too few cases to show a clinical impact of the described planes, their application in a large systematic study including sufficient numbers may show the clinical effectiveness of including these planes. We have shown here that it is feasible to include the evaluation of these planes in fetal cardiac scanning, to augment the comprehensive evaluation of the fetal heart.

Ancillary