Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
Recognition of prenatal-onset skeletal dysplasias has improved with advances in ultrasound imaging. Skeletal abnormalities can be recognized by two-dimensional (2D) ultrasound, but generating a precise diagnosis can be challenging. We aimed to determine whether three-dimensional (3D) imaging conferred any advantages over 2D imaging in these cases.
We studied five women with fetuses of 16–28 gestational weeks referred for abnormal ultrasound skeletal findings. First 2D and then 3D sonography was performed and the results compared.
The pregnancies resulted in the following skeletal dysplasias: thanatophoric dysplasia, achondrogenesis II/hypochondrogenesis, achondroplasia, chondrodysplasia punctata (rhizomelic form) and Apert's syndrome. For all five fetuses, the correct diagnosis was made in the prenatal period by analysis of the 2D images. In each case the 3D images confirmed the preliminary diagnosis and for many findings it improved the visualization of the abnormalities.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
The detection of prenatal-onset skeletal dysplasias has improved tremendously with the improvements made in two-dimensional (2D) imaging. However, generating a precise diagnosis can be challenging and there are many underlying reasons for this. First, there are over 150 distinct skeletal dysplasias and many of them are exceedingly rare1. Therefore, few centers have had the necessary experience and/or expertise to determine the exact diagnosis. Secondly, many of the clinical and radiological findings that delineate these skeletal dysplasias are not well visualized by 2D imaging. Lastly, the gestational age of the fetus at the time of imaging can influence the degree and severity of the dysplasia. Although an accurate diagnosis can be difficult to achieve, it is of enormous benefit to the family and consulting physicians if management can be based on a precise diagnosis. The most immediate concern for the family and physician should be whether or not the presenting skeletal dysplasia is predominantly lethal. However, an accurate diagnosis is also critical for the appropriate management of the fetus and to provide counseling to the family regarding future recurrences.
In addition to improvements in 2D ultrasound, new and rapidly available advances have been made in three-dimensional (3D) imaging. It is performed in a similar manner to that of 2D imaging though the saved images are simultaneously obtained in three planes, as opposed to two. These multiplanar images can then be reconstructed to create a 3D image. Several authors have advocated the value of 3D imaging in assessing the fetal skeleton, cranium and facies in both normal and abnormal gestations2–9. However, 3D ultrasound imaging is not routinely performed in many centers and its benefit over 2D imaging has not been firmly established. As with any new technology, studies need to be performed to determine whether the new methodology has proven benefit over the well-established standard methods. The purpose of this study was to evaluate fetuses with different skeletal dysplasias and compare the prenatal 2D with 3D ultrasound findings to determine if 3D imaging was superior.
Patients and methods
Five women were referred to the Medical Genetics Birth Defects Center/International Skeletal Dysplasia Registry because of abnormal ultrasound skeletal findings, though none had a definitive diagnosis. The gestational ages ranged from 16–28 weeks. None of the families was known to be at risk for offspring with skeletal dysplasias. We initially evaluated the fetuses using the HDI 5 (ATL/Philips, Bothell, WA, USA) 2D ultrasound machine with both 3.5- and 5.0-MHz transducers. During 2D ultrasound imaging, findings consistent with skeletal dysplasias were noted in all five gestations. Following completion of the 2D sonography, preliminary diagnoses were made based on the findings detected. The patients were then recruited for 3D ultrasound studies to determine whether data obtained by 3D imaging correlated to the 2D findings, and if 3D imaging offered any diagnostic superiority to 2D ultrasound. For each modality, the patient underwent ultrasound examination for approximately 60 min.
The 3D ultrasound data were saved to permit further evaluation and manipulation once the patients had left the clinic. The data were reviewed to identify ultrasound findings consistent with skeletal dysplasias and to analyze them relative to the 2D findings. The volume sets were rotated and the images were set up in standard anatomical orientation, with the fetuses in the upright position, and displayed as simultaneous topographic images in three planes (transverse, sagittal and coronal). A truncated pyramid demonstrated the plane of the section. The cranium and facies were evaluated subjectively for facial dysmorphisms. The limbs were evaluated for rhizomelia, mesomelia and acromelia, bowing, and mineralization patterns. Hands and feet were evaluated for relative size, posturing and configuration of the phalanges. The axial skeleton was inspected subjectively for mineralization patterns, platyspondyly and rib shape and length. The images were obtained using skeletal and surface rendering techniques. The 3D images were compared with the 2D images by two individuals who performed the sonography and the findings were scored and analyzed to determine whether the same, additional, or conflicting information was obtained. The information regarding the 2D and 3D findings were given to the referring physician and the patients were then counseled. All women opted to undergo termination of pregnancy.
To establish a definitive diagnosis in each case, the fetus underwent extensive postmortem evaluation that included dysmorphology examination, radiographs, cartilage histomorphology, and DNA diagnosis when applicable.
Each of the five fetuses evaluated had one of the following skeletal dysplasias: thanatophoric dysplasia, achondrogenesis II/hypochondrogenesis, achondroplasia, chondrodysplasia punctata (rhizomelic form) and Apert's syndrome. For all five fetuses, the correct diagnosis was made in the prenatal period by analysis of the 2D images. In each case the 3D images confirmed the preliminary diagnosis and for many findings it improved the visualization of the abnormalities.
The abnormal skeletal ultrasound findings seen by both 2D and 3D imaging are summarized in Table 1. These findings were subjectively scored by two individuals to determine whether the findings were better visualized by 2D and/or 3D ultrasound. As noted in the table, all of the ultrasound findings were identified using both modalities; however, 3D ultrasound was superior in elucidating some of the features. The findings obtained by 3D ultrasound were characteristic for the dysmorphic findings typical of each skeletal dysplasia. For example, the Binder facies (depressed nasal bridge, mid-face hypoplasia, small nose with upturned alae) is part of the spectrum of chondrodysplasia punctata (rhizomelic form) and is illustrated in Figure 1. Facial dysmorphisms such as frontal (metopic) bossing and mid-face hypoplasia found in achondroplasia and thanatophoric dysplasia are demonstrated in Figures 2 and 3, respectively. The facial dysmorphisms were not ascertained nearly as well by 2D ultrasound, though the flattened nasal bridges were seen on 2D sagittal images of the face.
Table 1. Scorings from comparison of two-dimensional (2D) and three-dimensional (3D) ultrasound findings in the five fetuses with skeletal dysplasias
Skeletal abnormalities detected by both 2D and 3D imaging were scored for improved visualization based on ultrasound modality. + indicated that the abnormality was seen; ++ indicated that the abnormality was seen more clearly; +++ indicated the abnormality was seen very clearly.
Rib splaying (metaphyseal)
On 2D imaging, the 22-week fetus with achondroplasia showed relative macrocephaly and rhizo-, meso- and acromelia both subjectively and by absolute long-bone measurements. Brachydactyly was seen on 2D images (Figure 4); however, 3D imaging captured the trident configuration of the hands (Figure 5), almost pathognomonic for achondroplasia. Furthermore, the observation of disproportion was far superior utilizing 3D imaging. Figure 2 (achondroplasia) illustrates this well. The left arm raised to the fetal forehead showed foreshortening of the rhizomelic (proximal) portion of the arm since the elbow should be seen at the level of the nose, not below the chin. In addition, the fetal hand was flexed and the distal digits extended to the proximal interphalangeal joints, while they should extend almost to the thenar eminence. The sagittal image of almost the entire fetus (Figure 6) with achondrogenesis II/hypochondrogenesis also revealed disproportion which aided in establishing the diagnosis. The very short arms, relative macrocephaly, almost absent neck and small compacted chest and abdomen illustrated this disorder well, and the image was quite similar to previously published photographs of this disorder10. The abnormal ultrasound findings of platyspondyly and small thorax were seen equally well in 2D and 3D ultrasound, and 3D ultrasound did not convey any advantages for these features, though this finding was only seen in two cases (thanatophoric dysplasia and hypochondrogenesis). Rib shortening was seen by both modalities, and in our cohort of cases, 3D imaging did not convey any additional information. The very rare finding of laryngeal stippling (Figure 7) visualized in the case of chondrodysplasia punctata (rhizomelic form) was preferentially seen by 3D imaging because of the ability to rotate the image 180° and we visualized multiple punctate calcifications throughout the larynx. In the 2D sagittal image of the neck, the abnormal calcifications were detected, but the extent of them was not. The laryngeal calcifications were confirmed by both postmortem radiographs and autopsy.
In this study we had the opportunity to compare the ultrasound images seen by 2D and 3D sonography in five fetuses with skeletal dysplasias. In addition, we had the opportunity to confirm our prenatal diagnosis by fetal dysmorphology examination, radiographs, histomorphology and molecular diagnosis when applicable. This is critical since the final diagnosis of a skeletal dysplasia should be made by radiographs and histomorphology, not by ultrasound images alone. In all five cases, the preliminary prenatal diagnosis was correct. This aided us in our understanding of which 2D and 3D ultrasound findings could be clinically and radiographically confirmed.
The imaging of the fetuses for dysmorphic features was truly enhanced by the employment of 3D imaging. While abnormal profiles were appreciated by conventional 2D imaging, the extent of the dysmorphic features was far better demonstrated by 3D imaging. This supports the work of others who have illustrated the usefulness of 3D imaging for the fetal facies5, 7. Surface rendering of the fetal facies allowed for evaluation of the metopic prominence contour, bony structure, nasal contour and overall relationship of facial features. The fetal limbs, especially the hands and feet, were well delineated by 3D ultrasound. The extent of brachydactyly was more fully appreciated by 3D than by 2D ultrasound. This is of clinical importance since measurements of the phalanges, palms and feet are rarely done by 2D ultrasound imaging and brachydactyly is under-appreciated. Brachydactyly as seen in achondroplasia, by the relative relationship of flexed phalanges to the palm, was well visualized using 3D imaging (Figure 2). While absolute long-bone measurements and centiles are easily obtained by conventional 2D ultrasound, 3D imaging had the significant advantage of showing the relative disproportion of limb segments seen in the skeletal dysplasias. Rhizo-, meso- and acromelia were easily delineated when evaluating the fetuses with achondroplasia, hypochondrogenesis and thanatophoric dysplasia.
Platyspondyly and small thorax appeared to be equally well visualized by 2D and 3D imaging. Platyspondyly is a very subjective parameter on ultrasound and is operator-dependent. Because it is subjective, we did not categorize either method as superior for this parameter. Two-dimensional ultrasound thorax circumference tables are available and helpful for determining lethality, especially if the chest is < 5th centile11. Therefore, though the small thoraces were appreciated by 3D imaging, we did not find that 3D imaging offered any value over 2D imaging for this parameter.
There were two distinct advantages of 3D over 2D imaging in these five cases. The ability to store the images and rotate the planes aided us. This was especially true in the case of chondrodysplasia punctata (rhizomelic form), because subsequent rotation of the images allowed us to delineate the extent of abnormal calcifications in the larynx, proximal humerus and femur joints. Moreover, the 3D images, especially of the facies and extremities, were easier for the patient to interpret. Rendered images of the abnormal facies or extremities were compared with clinical images and shown to the patients. The patients visualized the 3D images much better than they did the 2D images. This aided in genetic counseling. The patients had more confidence in the preliminary diagnosis presented to them, which helped in their decision-making.
Skeletal dysplasias occur rarely, approximately 1/400012. However, many of them present in the prenatal period, are associated with lethality and have significant recurrence risks. Therefore, accurate detection and diagnosis in the prenatal period is of value to the patient. Our findings in this study show that for many of the findings seen in the skeletal dysplasias, 3D imaging is superior to 2D imaging and can improve diagnostic accuracy.
We would like to thank the referring physicians and families for their participation. This work was supported by NIH HD2257 to D.L.R. and D.K., and a grant from the Medison Corporation to L.D.P.