Portal de Periódicos da CAPES

Newer imaging modalities in the prenatal diagnosis of skeletal dysplasias

2004; Wiley; Volume: 24; Issue: 2 Linguagem: Inglês

10.1002/uog.1712

1469-0705

Luís F. Gonçalves, Jimmy Espinoza, M. Mazor, Roberto Romero,

Congenital Diaphragmatic Hernia Studies

Skeletal dysplasias are a group of heterogeneous disorders affecting growth and morphology of the various segments of the skeleton1-4, with an estimated prevalence of 2–5/10 000 at birth5. Despite recent advances in imaging modalities and molecular genetics6, accurate prenatal diagnosis of skeletal dysplasias remains a clinical challenge7. The difficulties in prenatal diagnosis result from the large number of disorders, their phenotypic variability with overlapping features and the lack of a precise molecular diagnosis for many disorders8. Over the past 30 years, the classification of skeletal dysplasias has evolved from one based upon clinical/radiological/pathological descriptions to one that also includes the underlying molecular abnormality for a few conditions in which the defect is known9. The last version of the International Nosology and Classification of Constitutional Disorders of Bone was published in 200210 and included approximately 300 conditions, of which approximately 50 are apparent and identifiable at birth11. These conditions are relevant to the maternal–fetal medicine specialist as they constitute the group which may be detected before birth. Prenatal diagnosis is easier in the presence of a positive family history and a precise description of the phenotype since many disorders are inherited as autosomal dominant or recessive disorders6. However, it is not unusual that skeletal dysplasias are first suspected during routine sonographic examination after a shortened long bone or other abnormal skeletal finding is observed7. Sonographic evaluation requires a detailed examination of the fetal skeleton, which is described in Table 1. The sonologist must determine if the anomaly is lethal, either by establishing the precise diagnosis of uniformly lethal conditions (Table 2)6 or assessing the risk of pulmonary hypoplasia6, 12-35. Several parameters have been proposed to estimate the risk of pulmonary hypoplasia (see Table 3)6, 12-47. 2.Compare with other segments and classify the limb shortening as: Rhizomelia Mesomelia Acromelia Severe micromelia 3.Qualitative assessment of long bones: Bowing Demineralization Fractures Metaphyseal flaring Absence of bones 5.Evaluate hands and feet Digits (polydactyly/syndactyly) Positional deformities 6.Evaluate the cranium Macrocrania Frontal bossing Cloverleaf skull Hypertelorism/hypotelorism 8.Examination of the spine Platyspondyly Demineralization Hemivertebrae Coronal clefts Vertebral disorganization With the identification of a growing number of mutations responsible for skeletal dysplasias48-50, it is hoped that eventually a specific diagnosis will be possible during the prenatal period51-55. A classification of genetic disorders of the skeleton based on the structure and function of implicated genes and proteins has been proposed recently50, complementing the existing International Nosology and Classification of Constitutional Disorders of Bone10. Accurate diagnosis is important for both proper management of the index case and for genetic counseling56. Moreover, if patients resort to pregnancy termination, molecular diagnostic methods may be the only way to establish the final diagnosis5. For detailed and up-to-date information regarding molecular tests available for the diagnosis of skeletal dysplasias the reader is referred to the web pages of GeneTests (a comprehensive source of information on genetic testing funded by the National Institutes of Health of the USA: http://genetests.org), the European Skeletal Dysplasia Network (http://www.esdn.org/diagnosis.html), the Division of Molecular Pediatrics of the University of Lausanne, Switzerland (http://www.pediatrics.ch/Frames.html) or the International Skeletal Dysplasia Registry at Cedars-Sinai Hospital, Los Angeles, USA (http://www.cedars-sinai.edu/3810.html). The reader is referred to a recent novel article describing how phenotypic relationships could be exploited to find new disease genes and provide clues to gene interactions, pathways and functions57. Despite the increasing availability of molecular testing, a comprehensive molecular diagnostic search for all skeletal dysplasias is not possible at this time. Therefore, the role of imaging is to narrow the differential diagnosis so that specific biochemical and molecular tests can be performed to confirm or exclude a potential diagnosis8, 58-60. Ultrasound is the primary imaging modality used for the initial diagnostic evaluation of an affected fetus. Table 4 summarizes the diagnostic accuracy of two-dimensional ultrasound (2D-US) for prenatal diagnosis of skeletal dysplasias7, 61-64. A precise diagnosis has been reported in 31–65% of cases. The introduction of 3D-US and rendering algorithms to reconstruct the fetal skeleton can improve diagnostic accuracy because additional phenotypic features not detectable by 2D-US may be identified65-77. For example, Garjian et al.68 and Krakow et al.75 reported the diagnosis of additional facial68, 75, scapular anomalies68 and abnormal calcification patterns75 in fetuses with skeletal dysplasias, whereas Moeglin and Benoit71 used the multiplanar visualization method to demonstrate the ‘pointed appearance’ of the upper femoral diaphysis in a case of achondroplasia. Three-dimensional reconstruction of the fetal bones is best performed using the ‘maximum intensity projection’ mode, a rendering algorithm that prioritizes the display of the highest gray levels contained within a region of interest selected by the operator68, 71. If the fetus is examined early enough during pregnancy, the entire skeleton can be included within the region of interest of the three-dimensional volume dataset and, therefore, a panoramic visualization can be achieved68. However, the diagnosis may be missed, as the phenotypic characteristics of some skeletal dysplasias do not manifest until later in pregnancy. Case reports and small series of several skeletal dysplasias have been published describing phenotypic characteristics or skeletal features that were best demonstrated by 3D-US (Table 5)66, 68-71, 73-76. In this issue of the Journal, Ruano et al.78 explore for the first time the use of three-dimensional helical computerized tomography (3D-HCT) as an adjunctive imaging modality for prenatal diagnosis of skeletal dysplasias in six cases: achondroplasia (n = 3), osteogenesis imperfecta (n = 2), and chondrodysplasia punctata (n = 1). This technique consists of fast and continuous acquisition of images as the X-ray source and detector perform ‘spiral’ or ‘helical’ movements with respect to the patient, who is moving in a linear fashion through the gantry79, 80. Helical CT allows fast and continuous acquisition of a volume dataset containing the structures of interest, usually within one breathhold (20–25 s), minimizing motion artifacts79-82. Fast acquisition allows improvement in multiplanar reformatting and three-dimensional reconstruction capabilities over what is feasible with conventional CT. Similarly to 3D-US, postprocessing techniques such as ‘maximum intensity projection’, ‘surface rendering’ and ‘volume rendering’ can be used for three-dimensional reconstruction79, 81, 82. In addition, faster image acquisition with 3D-HCT when compared with conventional CT has the potential to decrease radiation exposure80-82. 3D-HCT has been previously used as an adjunctive diagnostic imaging modality for prenatal diagnosis of congenital anomalies in isolated cases of trisomy 18, cystic hygroma, congenital diaphragmatic hernia and agnathia-holoprosencephaly80, 83, 84. Long bone measurements obtained by postmortem helical CT studies have been compared with those obtained within 24 h of delivery by ultrasound, and a significant correlation between the two methods was observed85. In the study of Ruano et al.78, excellent panoramic images of the fetal skeleton were obtained by 3D-HCT without superimposition of the maternal skeleton, which occurs with radiography. Deformation of the fetal pelvis and an increase in the intervertebral space of the lumbar vertebrae were diagnosed more often using 3D-HCT when compared with 2D-US and 3D-US. In contrast, some phenotypic characteristics of fetuses with skeletal dysplasias were demonstrated only by ultrasound: phalangeal hypoplasia (by both 2D-US and 3D-US), facial dysmorphism (by 3D-US only) and point-calcified epiphysis (by both 2D-US and 3D-US). The overall count of correct phenotypic characteristics detected prenatally favored 3D-HCT over 3D-US (94.3% (33/35) vs. 77.1% (27/35), P = 0.03, McNemar's test for correlated samples). Of interest, however, was the observation that the diagnostic performance of 3D-HCT was not superior to that of 3D-US, as the correct prenatal diagnosis was established by both modalities in all cases. Provided that the two diagnostic methods have comparable diagnostic accuracy, 3D-US has two important advantages over 3D-HCT, namely, lack of radiation exposure and wider availability in the clinical setting. It is also noteworthy that the overall experience with 3D-US for the diagnosis of skeletal dysplasias is still limited65-77. Nevertheless, even in this study, 3D-US performed better than did 2D-US, both in the identification of phenotypic characteristics (77.1% (27/35) vs. 51.4% (18/35), P = 0.004, McNemar's test for correlated samples) and in establishing an accurate diagnosis. Ruano et al.78 proposed that the improved performance in detection of some features of skeletal dysplasias by 3D-HCT when compared with ultrasound could be attributed to 3D-HCT images being examined by both a radiologist and a geneticist, whereas ultrasound examinations were performed and analyzed essentially by a sonologist. It remains to be determined if the type of technology, the team approach, the difference in clinical skills, or a combination of all these factors, are ultimately responsible for the differences observed. Regardless of the cause, one way to minimize biases in the interpretation of cases of skeletal dysplasias examined by ultrasound would be to take advantage of 3D-US and information technology (e.g. acquisition and storage of volumes and the possibility of transmitting the files remotely using a fast Internet connection) to have cases re-examined at centers specialized in the diagnosis of skeletal dysplasias. This approach would maximize the diagnostic potential of 3D-US in the clinical setting by ensuring that physicians skilled in the differential diagnosis of skeletal dysplasias can examine the volume datasets. In addition, referral of cases to institutions appropriately equipped and with expertise in prenatal diagnosis would allow studies to be conducted to answer the very questions raised by the current study. It is expected that, as more prenatal diagnostic centers incorporate 3D-US into clinical practice, proficiency in the reconstruction of skeletal structures, as well as the diagnostic accuracy, are likely to improve. Before embracing the use of 3D-HCT, one must have a clear understanding of the goals of diagnostic imaging in the prenatal investigation of skeletal dysplasias: 1) to narrow the differential diagnosis of skeletal dysplasias so that appropriate confirmatory molecular tests can be selected; 2) to predict lethality; and 3) to identify the fetus with a skeletal dysplasia early enough in pregnancy so that the diagnostic workup can be completed before the limit of fetal viability. 3D-HCT will expose the fetus to some degree of radiation, with the potential for long-term harmful effects86, 87. As the number of centers using 3D-US increases, we look forward to the publication of larger series reporting on the accuracy of the combination of 2D-US and 3D-US for the diagnosis of skeletal dysplasias. In our opinion, ultrasound will remain the primary modality for evaluation of the fetus affected by skeletal dysplasias, and 3D-HCT could have a role in cases where the information provided by 3D-US is insufficient for counseling and management. The final diagnoses in the study of Ruano et al.78 were established by postnatal clinical and radiological findings. While this is the standard today, a definitive diagnosis can be established for a growing number of skeletal dysplasias, including the ones described in this study, using molecular and/or biochemical tests. For example, the differential diagnosis of achondroplasia at birth includes severe hypochondroplasia, cartilage hair hypoplasia (metaphyseal chondrodysplasia, McKusick type), and pseudoachondroplasia88. As more than 99% of the patients with achondroplasia have either a Gly380Arg substitution, resulting from a G to A point mutation at nucleotide 1138 of the fibroblast growth factor receptor 3 (FGFR3) gene, or a G to C point mutation at nucleotide 113889, a definitive molecular diagnosis is possible in the majority of cases90. Osteogenesis imperfecta type II is caused by mutations in either the COL1A1 or COL1A2 gene, resulting in abnormal molecular constitution of pro-collagen type I91-93. Diagnosis can be confirmed by biochemical analysis of collagen (collagen screening) and, if the results are equivocal, DNA sequencing of COL1A1 and COL1A294. Rhizomelic chondrodysplasia punctata (RCDP) is a genetically heterogeneous disorder, with three types (types 1, 2 and 3) and an extensive differential diagnosis, including X-linked recessive chondrodysplasia punctata 1 (CDPX1), warfarin embryopathy, X-linked dominant chondrodysplasia punctata (CDPX2), and chondrodysplasia punctata tibial-metacarpal type95. The most common form is type 1, which is caused by mutations in the PEX7 gene (L292X, G217R, and A218V)95-98. These three mutations can be detected in 60% of the cases, whereas 95% of the mutations can be identified by DNA sequencing of the entire PEX7 gene. RCDP types 2 and 3 are inherited in an autosomal recessive manner and are rarer than RCDP type 1. RCDP type 2 is caused by deficiency of the enzyme dihydroxyacetone phosphate acyltransferase, whereas type 3 is caused by deficiency of the peroxysomal enzyme alkyl-dihydroxyacetone phosphate synthase. In both cases, the diagnosis can be confirmed by measurement of the specific enzyme activity in cultured skin fibroblasts95. For a more detailed list of biochemical and molecular tests for the various forms of chondrodysplasia punctata, the reader is referred to the GeneReviews website (http://www.genetests.org/profiles/rcdp). Thus, we suggest that future studies of skeletal dysplasias by 3D-US or other diagnostic imaging modalities include a combination of imaging, histology and molecular techniques for final diagnosis75, whenever the molecular basis for the disorder is known.