Three‐dimensional ultrasound examination of the fetal central nervous system

Abstract

Three-dimensional (3D) ultrasound examination has been used to evaluate the fetal central nervous system (CNS)1-16. Different approaches have been suggested that may be employed in both basic and dedicated examinations. The purpose of this paper is to illustrate the practical details of the different techniques described so far. The technique used to obtain an ultrasound volume is adequately described by three elements: (1) the section that is used to start the acquisition of the volume (referred to in the following as the ‘start’ scan); (2) the angle of rotation of the mechanical sweep of the motorized probe; and (3) the quality of the acquisition that can be varied by the operator and depends on the number of sections obtained during the acquisition. Both the angle of rotation and the quality of the volume influence the acquisition time and, when this is too long, the probability of movement artifacts increases. Tailoring the size and quality of the volumes to the specific diagnostic requirements is important because it increases the efficiency of the scan. In the following we provide information for each of the applications discussed. The modalities for the analysis of ultrasound volumes have been described in depth previously4, 16, 17. The multiplanar mode is most frequently used for assessment of the fetal CNS6, 10, 13, 14. With this mode of display, the plane parallel to the acquisition plane or ‘start’ appears in the upper left corner of the screen and is identified with the letter A; the plane perpendicular to A but parallel to the ultrasound beam is identified with the letter B and appears in the upper right corner. The plane that is both perpendicular to the ‘start’ scan and the ultrasound beam is defined as C, and is frequently referred to as the coronal plane17 (Figure 1). The terminology may sound confusing at times. As discussed later, when dealing with the fetal brain, the coronal plane of the volume typically demonstrates a sagittal or axial section of the fetal head13. Schematic representation of multiplanar analysis of ultrasound volumes, with corresponding multiplanar analysis of the fetal face. A is the plane parallel to the acquisition or ‘start’ scan; B and C are the reconstructed orthogonal planes. The C-plane is also commonly referred to as the coronal plane. The arrows indicate the points that represent the intersection of the three planes. Although 3D ultrasound imaging can be used in many ways to evaluate the fetal (CNS), we have found that in practice there are mainly two useful applications: the multiplanar analysis of volumes obtained with an axial approach6, 13 and the multiplanar analysis of volumes obtained from a sagittal or coronal approach12, 14. 3D ultrasound examination can assist in evaluation of the spine18 and is particularly helpful in early neurosonographic studies1-3, 5, 19. It may also be used to improve the quality of two-dimensional (2D) images. These aspects are described separately below. One of the major difficulties with sonographic examination of the fetal anatomy is obtaining views that are not easily accessible. At mid-gestation, most fetuses are in a horizontal position, and transverse and coronal sections are usually easy to obtain. These scanning planes, however, have many limitations. While examining the fetal head, one of the most important views is probably the so-called median one20, which provides unique information on intracranial structures such as the corpus callosum and the cerebellar vermis. Unfortunately, this scanning plane is particularly difficult to obtain. Several approaches have been described, but they all require considerable ability, time and, frequently, a vaginal examination9. 3D ultrasound imaging allows visualization of a virtual median plane from a volume that has been acquired from an axial approach6, 13, 17. The ‘start’ scan is parallel to the skull base and demonstrates the cavum septi pellucidi, which is about halfway between the skull base and the parietal calvarium (Figure 2). In the second trimester of gestation, a mechanical sweep with an angle of rotation of 45° usually includes the entire cephalic pole. In later gestation, the angle has to be increased and an angle of 60° is usually necessary in the advanced third trimester. Static volumes acquired with low to medium quality are adequate and this helps to minimize movement artifacts. Schematic representation of acquisition of an ultrasound volume of the fetal brain by an axial approach, with corresponding ultrasound image. The transducer is positioned parallel to the fetal head at the level of the cavum septi pellucidi. The incident sound beam is at a right angle to the midline echo. When the volume is displayed in multiplanar mode (Figure 3) and the C-plane is aligned along the midline echo demonstrated in the A- and B-planes, a median view of the brain is visualized (Figure 4)17. Multiplanar analysis of the ultrasound volume obtained from the ‘start’ scan displayed in Figure 2. The arrows indicate the direction of the rotation required to align the midline with the C-plane. A, B and C, orthogonal planes. Ultrasound images showing the same volume as that in Figure 3. Once the A- and B-planes have been rotated so that the midline is aligned with the C-plane, a median view of the brain is demonstrated, in which it is possible to recognize the comma-shaped anechoic complex formed by the combination of the corpus callosum (CC) and the cavum septi pellucidi (CSP), the area of the third ventricle (3v) and the cerebellar vermis. The brainstem and part of the cerebellar vermis are obscured by the acoustic shadow cast by the skull base (large arrow). A, B and C, orthogonal planes. We have previously demonstrated that there is a good correlation between median planes obtained directly with 2D ultrasound and those reconstructed with 3D ultrasound (3D median views)13. For example, measurements of structures such as the corpus callosum and the cerebellar vermis are closely related. However, 3D median views have a few shortcomings11, 13. First, they do not usually allow differentiation between the corpus callosum and the inferior cavum septi pellucidi. A single comma-shaped sonolucent structure is most frequently seen, outlined superiorly by an echogenic line that, comparing the different planes, can be identified as the lower edge of the midline echo. Only rarely, with optimal visualization conditions, is the corpus callosum seen, and it appears as a sonolucent stripe interposed between the inferior cavum septi pellucidi and the superior midline (Figure 5). Second, visualization of the posterior fossa is hampered by acoustic shadowing of the petrous ridges of the skull base (Figure 4). This shadow usually spares the cerebellum but obscures the brainstem, thus preventing assessment of the normal relationship between these structures. This problem can be minimized, although not completely resolved, by keeping an angle of about 45° between the incident ultrasound beam and the midline when the volume is acquired (Figure 6). We have found that simultaneous representation of the cerebellum in the three orthogonal planes is valuable, as a comparison between sagittal and coronal planes is required to assess the presence and integrity of the vermis; this can be achieved easily by properly positioning the three orthogonal planes (Figure 7). By slicing the volume obtained with axial planes in different directions, it is also possible to obtain other views of the brain, such as the oblique-1 view20, which demonstrates the three horns of the lateral ventricle, and a view of the lateral surface of the brain, which demonstrates the Sylvian fissure (Figure 8). However, these views can only be demonstrated from the hemisphere distal to the transducer, as the proximal one is obscured by noise and artifacts. Ultrasound images in (a) and (b) were obtained from an ultrasound volume of unusually high quality; in (b) the reconstructed median plane allows recognition of the corpus callosum as an anechoic stripe with a thin echogenic contour that separates it from the underlying cavum septi pellucidi (CSP). More frequently, the two structures are not clearly separated and generate a single anechoic comma-shaped image on top of the third ventricle (b and c). In both cases, the thick superior echogenic contour is formed by the lower portion of the midline echo. CC, corpus callosum. To avoid shadowing from the skull base and allow better visualization of the cerebellar vermis and brainstem, the volume is acquired by angling the transducer so that the incident sound beam (arrow) forms an angle of about 45° over the midline echo (line) (a). This allows better insonation of the posterior fossa, as shown in the corresponding reconstructed median plane (b); the cerebellar vermis and brainstem are now clearly outlined. 4v, fourth ventricle. Simultaneous demonstration of the main anatomical landmarks attesting to the integrity of the cerebellar vermis with multiplanar analysis of an ultrasound volume of the fetal brain acquired by an axial approach. In plane A the presence of the cerebellar vermis (CV) is shown by echogenic tissue separating the fourth ventricle (4v) from the cisterna magna; the slightly less echogenic cerebellar hemispheres (CH) are seen on both sides of the vermis. In plane C the fourth ventricle and the two main fissures of the cerebellar vermis (arrows) are demonstrated. The position and insertion of the tentorium cerebelli on the occipital bone are also well visualized. A, B and C, orthogonal planes. By adjusting the position of the C-plane in this volume obtained with an axial approach (a) it is possible to visualize sagittal views depicting the different portions of the lateral ventricle (b) and the Sylvian fissure (c). These images, however, can usually be obtained only from the hemisphere distal to the transducer. Noise and reverberations hinder the image of the proximal hemisphere (arrow). In our experience, diagnosis of midline anomalies of the fetal brain can always be made accurately using 3D median views13. Although the corpus callosum is not usually visualized directly, agenesis is consistently associated with an absent or small cavum septi pellucidi, which is easily identified in the 3D median view (Figure 9). The different anatomic elements that allow characterization of the clinical entities of the Dandy–Walker complex (integrity and rotation of the vermis, depth of the cisterna magna and insertion of the tentorium) are usually well appreciated (Figure 10). Although the brainstem is partially obscured by acoustic shadowing in 3D median views, the posterior contour is usually well visualized, particularly when angled insonation is used, as described previously (Figure 6). Therefore, the relationship between the position of the vermis and the brainstem, which is a key feature of the Dandy–Walker complex, is usually well appreciated13, 21, 22. Accurate diagnosis of the different clinical entities that form the spectrum of this condition (megacisterna magna, Blake's pouch cyst, vermian hypoplasia and Dandy–Walker malformation) has been reported by using the 3D median plane (Figure 10)13. Needless to say, the spectrum of vermian abnormalities that can be encountered is large, and experience with antenatal diagnosis is limited so far. The level of precision of the diagnosis remains unclear, and it is certainly possible that subtle pathologies may not be visualized accurately using the approach described. Three-dimensional median views demonstrating (a) a normal corpus callosum (arrow), (b) complete agenesis of the corpus callosum and (c) partial agenesis of the corpus callosum (arrow). Although this figure shows only the reconstructed median plane, diagnosis always requires a comparison with the other two orthogonal planes to ensure that the section is properly orientated. 3v, third ventricle. Three-dimensional median views comparing a normal posterior fossa (a) with different abnormalities, including: (b) megacisterna magna, which is a normally positioned and intact cerebellar vermis and a large cisterna magna (arrow); (c) Blake's pouch cyst, a superiorly rotated but seemingly intact cerebellar vermis (arrow); (d) vermian hypoplasia, a superiorly rotated and hypoplastic vermis (arrow); and (e) Dandy–Walker malformation, a superiorly rotated vermis (arrow) in association with superior displacement of the tentorium cerebelli. Although this figure shows only the reconstructed median plane, diagnosis always requires a comparison with the other two orthogonal planes to ensure that the cerebellar vermis is indeed present and that the scan is properly orientated. Coronal and sagittal views of the fetal head are best obtained by aligning the probe with the fontanelles and sutures of the upper calvarium. This approach allows the visualization of fine details of intracranial anatomy, particularly when using a high-frequency vaginal probe. This type of examination is also frequently referred to as a fetal neurosonogram and has many advantages over basic examinations performed with transabdominal axial views9, 20. The ‘start’ scan is aligned either with the sagittal or the coronal suture (Figure 11). A wide angle of rotation is usually necessary to include the entire fetal brain, in general 60–80°. As subtle anatomic details of the fetal brain are revealed when using a high-frequency probe, we recommend acquiring volumes with medium to high quality. It is frequently necessary to manipulate the head of the fetus to align the probe adequately, which requires the use of both hands. For this reason, we have found that the availability of a foot switch to command the acquisition of the volumes is of considerable help. When the volume is analyzed in the multiplanar mode the A- and B-planes demonstrate sagittal and coronal views of the fetal head, depending on the start plane used, whereas the C-plane demonstrates an axial view (Figure 11). The volume may be examined using sequential slices in the coronal, sagittal or axial planes (Figure 12). By adjusting the position of the volume, views that demonstrate specific details of the fetal brain can be obtained. For example, by tilting the coronal plane of the fetal head from one side to the other it is possible to obtain the oblique-1 plane that demonstrates almost the entire lateral ventricle in a single image (the so-called three horn view)14. Multiplanar analysis of an ultrasound volume of the fetal brain at mid-gestation obtained with a transvaginal approach from the sagittal plane. A, B and C, orthogonal planes. Multiple slices of the volume shown in Figure 11 orientated along the sagittal plane. Visualization of the fetal brain with this approach is much better than that achieved with the standard axial technique. However, some acoustic shadowing continues to occur. In general, visualization of the frontal and parietal area is favored because of the large acoustic window of the bregmatic fontanelle. More posteriorly, the sagittal suture is thin, and this results in bilateral shadows (Figure 13). When performing a 2D ultrasound examination in real time this problem can be overcome by tilting the transducer so as to visualize the two sides sequentially. When analyzing a volume, clear visualization of both occipital lobes is usually impossible. (a) Maximum mode ultrasound image of the superior fetal skull demonstrating the bregmatic fontanelle and sutures; (b and c) coronal and axial planes from the volume shown in Figures 11 and 12 demonstrating the distribution of acoustic shadowing from the bones of the calvarial convexity (arrows). The largest acoustic window is present in the most anterior portion of the brain due to the presence of the bregmatic fontanelle. When the bregmatic fontanelle or the sagittal suture is used as an acoustic window, the frontal and parietal areas of the brain are usually well visualized, including structures such the superior the convexity of the brain, the corpus callosum and lateral However, the posterior fossa remains in the and is visualized particularly when using a high-frequency probe. The is much the posterior fontanelle is used as an acoustic window (Figure (a) Maximum mode ultrasound image of the of the fetal head demonstrating the posterior fontanelle (arrow); (b) multiplanar analysis of an ultrasound volume of the fetal head obtained by transvaginal scanning using the posterior fontanelle as an acoustic The structures of the posterior fossa and the brainstem are seen clearly than with the standard approach in Figure 11). 4v, fourth A, B and C, orthogonal planes. When the fetal brain is with 2D ultrasound using sagittal and coronal planes the of anatomic is usually Visualization is not by 3D examination and there is the of acoustic The of 3D ultrasound imaging in such in the at which a volume is thus analysis at that is not by fetal in the volume the of structures by comparing the three orthogonal planes, and the and of 14. The combination of 3D ultrasound with imaging has been used to demonstrate the anatomic details of complex of the fetal (Figure More visualization of the entire has been Ultrasound images of an of the of (a) image obtained by an axial approach demonstrating the (arrow); (b) transvaginal image demonstrating that the of not the but superiorly to the and the superior sagittal (c) of an ultrasound volume obtained from a transvaginal approach demonstrating the different elements of the in a single including the between the of the skull base (arrow) and the of which the The use of 3D examination of the fetal has been in One of the main advantages is the of the entire of the elements of the of the fetus in one single image. The ‘start’ scan is a view of the and the best results are obtained when the transducer is aligned with the plane of the from the of the It is to include the in the view as and, in general, an angle of rotation of is As the elements are small and of the fetus frequently generate good results are also obtained with ultrasound which volumes at high The volume can be displayed either in multiplanar mode or in which demonstrates the of each (Figure The images thus obtained are useful for the assessment of such as agenesis and (Figure We have however, that a normal 3D image of the elements of the is not with to the presence of is associated with a in the of the of the lateral of the posterior At however, the can be and difficult to particularly in the of small and low (Figure it is a that we have also in fetuses with anomalies but with an intact of the volume can be used to demonstrate the of and the possible presence of a (Figure but this may be and diagnostic over standard 2D ultrasound One of 3D ultrasound imaging in the multiplanar mode is to the and position of a When using standard 2D ultrasound the area of the is identified by obtaining a view and the from the corresponding to With multiplanar 3D ultrasound the are with the one to the this usually to (Figure Three-dimensional ultrasound images of the fetal The best results are obtained by using the sagittal view as a ‘start’ scan (a). In multiplanar by the C-plane from posterior to anterior as by the in it is possible to demonstrate the (arrows) of the lateral (b) and of the (c). mode in a thin it is possible to visualize the three (arrows) of the (a) Maximum mode of fetal (arrow) as a and an of the (b) agenesis (arrow); and a were also The of the and is well demonstrated by two-dimensional ultrasound examination (arrow) but ultrasound results in only subtle and (b and c). In multiplanar an increased of the lateral is in the area and it is also possible to visualize an between the lateral in the area the (b); in mode the are subtle (c). The of (a) is with surface of antenatal ultrasound volumes (b and (d) surface of The arrows indicate the associated with the of the level of using In the sagittal plane the the intersection of the orthogonal planes (arrow) is positioned at the upper level of the which is well visualized in this in the coronal plane the number of elements between the corresponding to and the by the intersection of orthogonal planes (arrow), is in this of the A, B and C, orthogonal planes. 3D ultrasound examination has been used in of fetal 5, and has also the early diagnosis of including and neurosonographic examinations are usually performed with transvaginal that are limited in of which frequently results in the to obtain views adequate for The of 3D scanning is that obtaining a volume of the fetal cephalic is usually and this can be along planes later (Figure At this time in gestation, the fetal is not and so the acoustic window is In our examination of the brain in the trimester much and when 3D than 2D ultrasound examination examination of an at and Although the position of the is not for analysis of the brain a volume can be obtained and the aligned along a Multiplanar analysis (b) allows clear of subtle anatomical details 5, of A, B and C, orthogonal planes. imaging is an of 3D ultrasound that a thin from an acquired One of the main of this technique is to ultrasound artifacts. By and different of and noise that are at are or anatomic structures are (Figure This results in an image with noise and It has to be demonstrated increases the of information over that by standard multiplanar However, the images to be and the between different is to the can be used as a technique for volumes that have been acquired or as a for in mode in the coronal We have with a of as the of to demonstrate multiplanar analysis of the fetal all the demonstrated in this have been obtained by using this is also (Figure This allows of the median plane in almost real the main is of movement artifacts that are encountered when acquiring has been employed to of cerebellar and has been found to be as as 3D imaging in median planes from transverse We however, our for a volume of medium quality of the fetal head can be obtained and has the of better in the three orthogonal planes. This is particularly useful when dealing with complex The main of in the fetus in our experience so has been in the demonstration of the skull bones (Figure and fetal (Figure In this the is in the (a) Schematic representation of the of volume imaging the of different of tissue allows a better The of is demonstrated by comparing images obtained from the same ultrasound volume, displayed using a standard mode (b) and with with a of (c). volume imaging of the fetal brain (a) and fetal CC, corpus callosum. The ‘start’ are on the left and the coronal planes are on the 3D ultrasound imaging is used in examinations. It remains to be demonstrated this have a on the diagnosis of fetal However, there is a general that some at in of fetal examination and At 2D ultrasound examination remains the of fetal diagnosis for many However, in the specific of sonographic evaluation of the fetal 3D imaging is helpful in that it allows acquisition of scanning planes along different from that of the incident sound It has been that 3D ultrasound imaging not overcome the of ultrasound and that sections reconstructed from ultrasound volumes are less than those obtained by images always be with 13. we that fetal performed by to mainly of 2D images as these allow of anatomic which may be particularly when dealing with However, reconstructed planes are useful for the assessment of normal anatomy in standard examinations as well as for assessment of when be examination of the fetal brain is an in this mainly in the the fetal is not an to ultrasound and images that are obtained by volumes are of almost the same quality as those obtained We that 3D ultrasound examination has many advantages a neurosonogram is applications of this have a less on evaluation of the fetal However, the of ultrasound volumes in of analysis and not be The following is from the the images from this with This the from the The is not for the or of information by the than be to the corresponding for the