Chapter 19: FETAL SKELETAL ANOMALIES

The human skeleton (from the Greek word skeletos, dried up) is a complex organ that consists of 206 bones. It has multiple embryonic origins and serves many key functions, including mechanical support for movement, protection of vital organs, and a reservoir of blood and minerals.¹ Fetal skeletal dysplasias are a heterogeneous group of disorders that result in abnormalities in the size and shape of various segments of the skeleton due to intrinsic derangement of the growth, development, and/or differentiation. Nearly 300 disorders are included in this entity; most of them are heritable diseases.² Although most of the diseases are rare, skeletal dysplasia as an entity is common. There are two classes of skeletal dysplasia: osteochondrodysplasia and dysostosis.² Osteochondrodysplasia results from abnormal growth and development of bone and/or cartilage; the lesions are generalized and progressive. Examples include achondroplasia and osteogenesis imperfecta. In contrast, dysostosis is the disorder of individual bones, singly or in combination; the lesions are local and nonprogressive. Poly/syndactyly and craniosynostoses are representative diseases. Previously, it was thought that the two were different in that osteochondrodysplasia was considered a genetic disorder, whereas dysostosis was an accidental derangement during embryonic development (unrelated to genes).² However, it is now known that both are caused by genetic abnormalities. Responsible genes for skeletal dysplasia have been identified in more than 150 diseases mainly through positional cloning.² In recent years, major advances have been made in identifying mutations associated with chondrodysplasias. This chapter reviews progress that has been made in this field, the birth prevalence and classification of skeletal dysplasias, and provides an approach to the diagnosis of these conditions in utero.

The process of skeletal formation and growth includes the differentiation of mesenchymal cells to form cartilage for future bone endochondral ossification. The growth of the long bones occurs through differentiation of chondrocytes in the growth plates and intramembranous ossification.^3,4 Disruption in any of these processes results in skeletal abnormalities.⁵ Although a wide range of phenotypes has been described in osteochondrodysplasias, recent advances in the understanding of the molecular basis of skeletal dysplasias indicate that a spectrum of phenotypes share a similar genetic basis.^6,7 Although the familial tendency of chondrodysplasias has been known for many years, the molecular basis for a number of conditions has only recently been clarified (Table 19-1).^8,9

Bone Morphogenetic Proteins

Bone morphogenetic proteins (BMPs) are products of a family of genes that play a role in regulating the early mesenchymal cell condensation into the chondrogenic and osteogenic cell lineages, and in influencing the size and shape of skeletal elements.^10,11 BMPs have the ability to induce ectopic bone formation when implanted subcutaneously or in muscle.^12,13 BMPs have homology with the superfamily of transforming growth factors-β.^{14, 15, and 16} Mutations in the cartilage-derived morphogenic protein-1 (CDMP-1) have been shown to cause acromesomelic chondrodysplasias (Hunter-Thompson type), and autosomal dominant brachydactyly type C (see Table 19-1).^{17, 18, 19, and 20}

Mutation in Transcription Factor Genes

Transcription factors are a class of proteins that play an important role in regulatory gene expression. They bind to regulatory elements in DNA promoters or enhancers that result in stimulation or inhibition of gene transcription and mRNA formation.^{21, 22, and 23} They are arranged into distinct domains, and are important for the development of the embryo. Mutations in the genes coding for this group of proteins may cause chondrodysplasias.

Homeotic Genes

Homeotic change is the transformation of one body part into another. Homeotic (HOX) genes control the morphogenesis of body parts. These genes encode for a family of proteins (transcription factors) that have a DNA-binding domain called the homeodomain or homeobox.^{24, 25, 26, 27, and 28} Transcription factors regulate gene expression by binding to DNA, and play a central role in skeletal pattern formation.^{28, 29, 30, 31, 32, and 33} Inactivation or pathologic expression of these genes causes deletion or addition of skeletal elements.^31,32 Homeobox genes located in the 5′ section of the HOXA and HOXD complexes are required for proliferation of skeletal progenitor cells of the limb. Specific combinations of gene products determine the length of the upper arm (genes belonging to groups 9 and 10), the lower arm (groups 10, 11, and 12), and the digits (groups 11, 12, and 13).³⁴ In humans, mutation in the HOXD-13 gene causes synpolydactyly (webbing and insertion of an extra digit) in heterozygous individuals.³⁵

PAX (paired like homeobox-containing) genes are a gene family that share a sequence called paired box, that codes for a protein domain with DNA-binding properties.^36,37 This transcription factor is also involved in skeletal formation. Mutation in PAX-3 has been implicated in Waardenburg syndrome, an autosomal dominant skeletal dysplasia with craniofacial anomalies.^{38, 39, 40, 41, and 42} Mutation in another transcription factor gene, MSX2, causes Boston type craniosynostosis.^43,44

SOX and TBX Genes

The gene responsible for campomelic dysplasia has been localized to chromosome 17.^{45, 46, and 47} A substantial fraction of XY patients with campomelic dysplasia (75%) demonstrate sex reversal syndrome. The development of the testis in mammals requires the sex-determining region Y (SRY) gene,^{48, 49, 50, and 51} and the protein encoded by this gene includes a high mobility group (HMG) domain, termed HMG box. This SRY HMG box binds to a specific DNA sequence that confers DNA transcription properties. Many other genes that encode proteins related to HMG boxes have been identified, and those that encode proteins with a similarity of greater than 60% to the SRY HMG region have been termed SOX (SRY box).^{52, 53, and 54} The SRY gene exerts its effect by interacting with the adjacent SOX-9 gene.⁵⁵ SOX-9 is required for cartilage formation,^56,57 and mutations of the SOX-9 gene have been reported in several patients,^58,59 indicating that both campomelic dysplasia and sex reversal can be caused by this mutation.^{58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, and 69}

Holt-Oram syndrome is characterized by cardiac septation defects and upper limb malformations. Linkage studies have mapped the gene responsible for this condition to chromosome 12q2. Mutations in one of the human T-box transcription factor family genes (TBX-5) underlie this skeletal disorder.^{70, 71, 72, 73, and 74}

Fibroblast Growth Factor Receptors

Fibroblast growth factors are a group of at least 9 heparin binding polypeptide growth factors that have pleotropic effects on many different cell types at various stages of development.^75,76 The cellular responses to these proteins are mediated through cell surface receptors.⁷⁷ Different genes encode for different, but highly homologous receptors. Fibroblast growth factor receptor (FGFR) mutations cause a spectrum of disorders linked to the degree of receptor activation.⁷⁸ Recurrent mutations of single amino acids in the transmembrane domain of FGFR3 causes a number of disorders. Achondroplasia, thanatophoric dysplasia, and hypochondroplasia are skeletal dysplasias resulting from mutations in FGFR3.^{79, 80, and 81} Graded activation of FGFR3 by mutations causes achondroplasia and thanatophoric dysplasia.⁸⁰ Achondroplasia, which is nonlethal, results from a glycine-to-arginine substitution at position 380 in the transmembrane domain of the tyrosine kinase receptor.⁸¹ Distinct mutations in FGFR3 cause two different types of thanatophoric dysplasia.⁸² On the other hand, mutations in FGFR1 and FGFR2 have been described in Pfeiffer syndrome,⁸³ whereas Apert syndrome,^{84, 85, 86, 87, and 88} Jackson-Weiss syndrome,⁸⁹ and Crouzon syndrome are associated with mutations in FGFR2.^{90, 91, 92, 93, 94, and 95}

Parathyroid Hormone Related Protein Receptor

Jansen metaphyseal chondrodysplasia is a short limb dwarfism caused by abnormal development of the growth plate. Hypercalcemia and hypophosphatemia with normal or low levels of parathyroid hormone (PTH) and PTH-related protein receptor (PTHrPR) characterize this disorder.⁹⁶ Schipani et al identified a mutation of the gene encoding for this receptor in a patient with Jansen metaphyseal chondrodysplasia.⁹⁷ The mutation results in a persistent activation of the receptor shared by these 2 hormones and leads to Jansen metaphyseal chondrodysplasia.^98,99 The mutation results in a change of a histidine to an arginine residue (H223R), causing ligand-independent activation of the receptor.^100,101 Recently, genomic DNA from an additional 6 patients with Jansen disease was analyzed and revealed a second activation mutation in the PTH/PTHrPR.¹⁰² This novel mutation causes a change of threonine to proline (T10P).^103,104

Blomstrand osteochondrodysplasia (BOCD) is a rare lethal skeletal dysplasia characterized by accelerated endochondral and intramembranous ossification.^105,106 A comparison of the characteristics of BOCD with type I PTH/PTHrPR-ablated mice show similarities that are most prominent in the growth plate. Therefore, this overall similarity suggested that an inactivating mutation of the PTH/PTHrPR might be the underlying genetic defect causing BOCD. Inactivating mutations of the PTH/PTHrPR have recently been identified.^{106, 107, 108, 109, 110, and 111}

Collagen

Collagen is the most abundant extracellular matrix protein. It is synthesized from 3 promolecules known as α chains.^112,113 Type 1 collagen is a heterotrimer of 1 (α1) and 2 (α2) chains, and is the most abundant protein in bone.¹¹⁴ Type 2 collagen, a homotrimer, is the major structural component of cartilage. It also has a role in other structures, such as the vitreous body, intervertebrate disc, and tectorial membrane of the inner ear.^{115, 116, and 117} Mutations in type 1 collagen gene locus α1 (COL1A1) and COL1A2 genes cause osteogenesis imperfecta.^{118, 119, 120, 121, 122, 123, 124, 125, and 126} The nature of the mutation and the consequent protein structure impacts the severity of the phenotype.¹²⁷ This impact decreases as the mutation shifts toward the amino terminal portion of the molecule. More than 30 mutations have been found at the type 2 collagen gene locus (COL2A1).^{128, 129, 130, 131, and 132} Mutations in other collagen types have been described with multiple epiphyseal dysplasia and Schmid metaphyseal chondrodysplasia.^{133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, and 146} Table 19-1 presents the important mutations, and the clinical phenotypes associated with them.

Cartilage Oligomeric Matrix Protein

Cartilage oligomeric matrix protein (COMP) is a noncollagenous component of cartilage matrix.¹⁴⁷ It is a pentamer of 5 identical subunits belonging to the thrombospondin family and is found in chondrocytes.^{148, 149, and 150} The gene for COMP was localized to chromosome 19. A mutation in the gene for COMP causes pseudoachondroplasia, an autosomal dominant chondrodysplasia.^{151, 152, 153, 154, 155, 156, 157, and 158}

Diastrophic dysplasia (DTD) is an autosomal recessive osteochondrodysplasia characterized by dwarfism,^159,160 spinal deformation,¹⁶¹ and joint abnormalities.^162,163 Undersulfatation of proteoglycans in cartilage matrix,¹⁶⁴ abnormalities in chondrocytes, and disorganization of collagen fibrils have been documented in this disorder.^{163, 164, 165, and 166} The disease is highly prevalent in Finland.¹⁶⁷ The gene encoding for the sulfate transport protein is designated DTDST (diastrophic dysplasia sulfate transporter) and is the same gene for DTD (chromosome 5q). Mutations in this gene lead to defective sulfatation of proteoglycans.^{168, 169, and 170} Mutations in the DTDST have also been found in achondrogenesis IB and atelosteogenesis type II.^{171, 172, 173, 174, 175, and 176}

The estimated birth prevalence of recognizable skeletal dysplasias, exclusive of limb amputations, is 2.4 per 10,000 births.¹⁷⁷ In a large series, 23% of affected infants were stillborn, and 32% died during the first week of life. The overall frequency of skeletal dysplasias among perinatal deaths was 9.1 per 1000. The birth prevalence of the different skeletal dysplasias and their relative frequency among perinatal deaths in this study are shown in Table 19-2. The 4 most common skeletal dysplasias found were thanatophoric dysplasia, achondroplasia, osteogenesis imperfecta, and achondrogenesis. Thanatophoric dysplasia and achondrogenesis accounted for 62% of all lethal skeletal dysplasias.¹⁷⁷ The most common nonlethal skeletal dysplasia was achondroplasia.

In another large series reporting the prevalence and classification of lethal neonatal skeletal dysplasias in western Scotland, the prevalence was 1.1 per 10,000 births, and the most frequently diagnosed conditions were thanatophoric dysplasia (1 per 42,000), osteogenesis imperfecta (1 per 56,000), chondrodysplasia punctata (1 per 84,000), campomelic syndrome (1 per 112,000), and achondrogenesis (1 per 112,000).¹⁷⁸ Rasmussen et al reported a prevalence of 2.14 cases per 10,000 deliveries in a longitudinal study that included elective pregnancy terminations, stillborn infants at more than 20 weeks of gestation, and liveborn infants diagnosed by the fifth day of life. The rate of lethal cases was 0.95 per 10,000 deliveries.¹⁷⁹ Table 19-3 shows the rate of skeletal dysplasias per 10,000 births in 11 studies.

Because there is great heterogeneity of skeletal dysplasias, classification is necessary for better comprehension and to provide practical diagnostic assistance. In order to develop a uniform terminology, a group of experts met in Paris in 1977 and proposed an International Nomenclature for Skeletal Dysplasias based on descriptive findings of either clinical or radiological nature.¹⁸⁰ This nomenclature underwent a revision in 1992,^{181, 182, and 183} and was reclassified using radiodiagnostic and morphologic criteria. The new classification grouped morphologically similar disorders into families of diseases based on presumed pathogenetic similarities. The International Working Group on Bone Dysplasias met in Los Angeles, California, in 1997 to update the Paris nomenclature of Constitutional Disorders of Bone.^184,185

In the revised nomenclature, families of disorders were rearranged based on recent etiopathogenetic information concerning the gene and/or protein defect. Disorders in which the defect is well documented were regrouped into distinct families based on the mutations. These include the "Achondroplasia group" of disorders with mutation in FGFR3, the "Diastrophic dysplasia group" of disorders with mutation in the DTDST gene, and the "Type II collagenopathies" with mutation in type II collagen. Several new groups were added, including the "Lethal skeletal dysplasias," the group with fragmented bones, and the "Miscellaneous neonatal severe dysplasia" group. Other groups were renamed (Table 19-4). Subsequently in 1999, the International Skeletal Dysplasia Society was established. In order to cope with the increasing complexity of information, a revision of the Nosology of Constitutional Disorders of Bone was performed in 2006, whose objective was to incorporate newly recognized disorders and reflect new molecular and pathogenetic concepts.¹⁸⁶ A table presenting this revised information may be found on the International Skeletal Dysplasia Society Web site (http://www.isds.ch/ISDSframes.html). A total of 372 different conditions were included and placed in 37 groups defined by molecular, biochemical, and/or radiographic criteria. Of these conditions, 215 were associated with one or more of 140 different genes. The number of recognized genetic disorders with a significant skeletal component is growing, and it seems that the distinction between dysplasias, metabolic bone disorders, dysotoses, and malformation syndromes is blurring.¹⁸⁶

Spranger and Maroteaux classified lethal osteochondrodysplasias in 11 groups, based on radial and anatomic manifestations (Table 19-5). The purpose of this classification is to facilitate the differential diagnosis, and the groups do not necessarily constitute pathogenetic "families."¹⁸⁷

This chapter focuses primarily on the osteochondrodysplasias that are recognizable at birth. Although approximately 300 skeletal dysplasias have been described and more will probably be identified as distinct entities, the number that can be recognized with the use of prenatal sonography is considerably smaller. Most of these disorders result in short stature; the term dwarfism has been used to refer to this clinical condition. However, in this chapter, we use the term dysplasia.

In the description of bone dysplasias, there is terminology that is commonly used. Shortening of the extremities can involve the entire limb (micromelia) (Figure 19-1), the proximal segment (rhizomelia) (Figure 19-2), the intermediate segment (mesomelia), or the distal segment (acromelia). The diagnosis of rhizomelia or mesomelia requires comparison of the dimensions of the bones of the legs and forearm with those of the thigh and arm. Figures 19-3, and 19-4 show the relationship between the humerus and ulna, and between the femur and tibia, respectively, and can be used in the assessment of rhizomelia and mesomelia. Table 19-6 presents skeletal dysplasias characterized by rhizomelia, mesomelia, acromelia, and micromelia.

Several skeletal dysplasias feature alterations of the hands and feet. The term polydactyly (Figure 19-5) refers to the presence of more than 5 digits. It is classified as postaxial if the extra digits are on the ulnar or fibular side, and as preaxial if they are located on the radial or tibial side. Syndactyly (Figure 19-6) refers to soft tissue or bony fusion of adjacent digits. Clinodactyly consists of permanent deviation of a finger (or fingers). Other descriptive terms used frequently involve the spine. The most common spinal abnormality seen in skeletal dysplasias is platyspondylia, which consists of flattening of the vertebral bodies (Figures 19-7 and 19-8). A normal spine for comparison is seen in Figure 19-9. Kyphosis and scoliosis can also be identified in utero (Figures 19-10, 19-11, and 19-12). The prenatal diagnoses of hemivertebra (Figure 19-13) and coronal clefting of vertebral bodies have also been reported.¹⁸⁸

In order to assess the normality of bone dimensions, gestational age is the independent variable, while the long bone is the dependent variable. For the proper use of these nomograms, however, the clinician must accurately know the gestational age of the fetus. Therefore, patients at risk for skeletal dysplasias should be advised to seek prenatal care at an early gestational age for dating purposes, and also to assess fetal biometry. Tables 19-7 and 19-8 present nomograms for the assessment of limb biometry for the upper and lower extremities, respectively. For those patients presenting with an uncertain gestational age, comparisons between limb dimensions and the head perimeter can be used (Figures 19-14 and 19-15). While some investigators have employed the biparietal diameter (BPD) for this purpose, the head perimeter has the advantage of being shape independent. A limitation of this approach, however, is that it assumes that the cranium is not involved in the dysplastic process, and this may not be the case in some skeletal dysplasias (eg, cloverleaf skull in thanatophoric dysplasia).

In this chapter, the nomograms and figures provide the mean, 5th percentile, and 95th percentile of limb biometric parameters. It is important to remember that 5% of the general population will fall outside these boundaries. Ideally, a more stringent criterion, such as the 1st percentile of limb growth for gestational age, should be used for diagnosis. Unfortunately, none of the currently available nomograms have been based on the number of patients required to provide an accurate discrimination between the 5th and the 1st percentiles. Despite this, most skeletal dysplasias diagnosed in utero or at birth are associated with dramatic shortening of the long bone(s), and therefore under these circumstances, the precise boundary used (1st or 5th percentile) is not critical. An exception to this is achondroplasia, where the limb biometry is only mildly affected until the third trimester. In this case, abnormal growth can be detected by examining the slope of femur length growth.¹⁸⁹

In a study including 17 skeletal dysplasias (127 cases), Gonçalves and Jeanty conducted discriminant analysis and showed that the femur length is the best biometric parameter to distinguish among the five most common disorders: thanatophoric dysplasia, osteogenesis imperfecta type II, achondrogenesis, achondroplasia, and hypochondroplasia.¹⁹⁰ Subsequently, Gabrielli et al¹⁹¹ evaluated the possibility of an early diagnosis of skeletal dysplasia in high-risk patients using transvaginal sonography. A total of 149 consecutive, uncomplicated singleton pregnancies at 9 to 13 weeks after amenorrhea were included in the study. The relationship between femur length and menstrual age, and between BPD and crown-rump length was established by using a polynomial regression. Eight patients with previous skeletal dysplasias were evaluated with serial examinations every 2 weeks from 10 to 11 weeks. A significant correlation between femur length and crown-rump length and BPD was found, while none was observed between femur length and menstrual age. Of the 5 cases with skeletal dysplasias, only 2 (1 with recurrent osteogenesis imperfecta and 1 with recurrent achondrogenesis) were diagnosed in the first trimester. This study concluded that an early evaluation of fetal morphology in conjunction with the use of biometric charts of femur length against crown-rump length, and femur length against BPD may be crucial for early diagnosis of severe skeletal dysplasias. However, by contrast, biometric evaluation appeared to be of no value for diagnosis in less severe cases.

Because of continued advancements and improvements in ultrasound technology, the potential for prenatal diagnosis of skeletal dysplasias continues to increase. However, the diagnosis of a specific skeletal dysplasia can remain difficult.

Several studies have explored the role of prenatal ultrasound in the detection of skeletal dysplasias.^{189,192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, and 203} Kurtz and Wapner performed a prospective analysis of a high-risk population (15 women, 16 cases) carrying a genetic risk for skeletal dysplasias.¹⁸⁹ Based on ultrasonographic findings in the second trimester, they were able to diagnose 5 (of 16) abnormal fetuses. In a study by Sharony et al,¹⁹⁵ a total of 226 fetuses and stillbirths referred for suspected skeletal dysplasia were evaluated. They reported an accurate prenatal diagnosis by the referring physician in less than one-third of cases. The most common final diagnoses were osteogenesis imperfecta (16%) and thanatophoric dysplasia (14%). Gaffney et al reported a series of 35 cases suspected for skeletal dysplasia. Using a systematic approach to ultrasound scanning in the second and third trimesters, they were able to predict the prognosis in 91% of cases, although an accurate diagnosis was only made in 31% of cases.¹⁹⁶ In contrast, in a large, high-risk pregnancy population (12,200) that was studied over 10 years with a wide range of gestational ages (9 to 40 weeks), there were 39 cases of skeletal dysplasia. Of these cases, 76.9% (n = 30) were detected in the first and second trimesters.²⁰⁴ Tretter et al²⁰⁵ identified 26 of 27 cases of lethal skeletal anomalies, and Hersh et al²⁰⁶ predicted lethality in 23 of 27 cases. The International Skeletal Dysplasia Registry was able to make a specific diagnosis in 81.5% of cases.²⁰⁷ This study also reported that the most likely time of diagnosis was between 18 and 20 weeks of gestation (consistent with the timing of the first routine ultrasound examination), with a further cluster in the third trimester as a result of investigation of specific pregnancy complications (ie, intrauterine growth restriction).

In a series of seven cases, the sonographic features of skeletal dysplasias obtained by two-dimensional (2D) and three-dimensional (3D) ultrasound were compared. Three-dimensional was better than 2D imaging in depicting abnormal spatial relationships, such as short ribs, splayed digits, and absent bones. In 3 of the 7 patients, 3D sonography provided additional information in the evaluation of skeletal dysplasias.²⁰⁸

The challenge of prenatal diagnosis of skeletal dysplasias generally presents itself in one of these 2 circumstances: (1) a patient has previously delivered an infant with a skeletal dysplasia and desires prenatal assessment of a subsequent pregnancy; or (2) there is an incidental finding of a shortened, bowed, or anomalous extremity during a routine sonographic examination. When patients are at risk, the examination is easier when the particular phenotype is known. An accurate history, including family history, is the foundation upon which the evaluation of any suspected skeletal dysplasia is built. Unexplained fetal deaths, consanguinity, and other family members with short stature, "orthopedic" problems, or "early arthritis" may provide valuable additional clues to diagnosis or possible modes of inheritance.¹ However, after identifying an incidental finding, the inability to obtain reliable information concerning skeletal mineralization and the involvement of other systems (eg, skin) with sonography is a limiting factor in the establishment of an accurate diagnosis. Another limitation is the paucity of information about the in utero natural history of these disorders.

Nevertheless, despite these difficulties and limitations, there are good medical reasons for attempting an accurate prenatal diagnosis of skeletal dysplasias. A number of these disorders are uniformly lethal, and a confident antenatal diagnosis would allow the patient to have the option of pregnancy termination. Table 19-9 lists the lethal skeletal dysplasias. Other skeletal dysplasias are associated with mental retardation,²⁰⁹ and this information would be important in prenatal counseling. In addition, there is another group of disorders which are associated with thrombocytopenia, and vaginal delivery may expose these infants to the risk of intracranial hemorrhage.

We propose the following systematic approach to the prenatal diagnosis of skeletal dysplasias.

Evaluation of Long Bones

All long bones should be measured in each extremity. In addition, comparisons with other segments should be performed to establish whether the limb shortening is predominantly rhizomelic, mesomelic, or acromelic, or whether all segments are affected (Figure 19-16). A detailed examination of each bone is necessary to exclude the absence or hypoplasia of individual bones (femur, tibia, fibula, humerus, radius, and ulna), which are frequently absent in certain conditions (Figures 19-17 and 19-18).^{192,210, 211, 212, and 213} In addition, the scapulae should be examined; for example, hypoplastic scapula can be seen in campomelic dysplasia.

An attempt should be made to characterize the degree of mineralization. This can be assessed by examining both the acoustic shadow behind the bone and the echogenicity of the bone itself. Signs of demineralization include visualization of an unusually prominent falx in the head, and the absent or decreased echogenicity of the spine. It should be stressed, however, that there are limitations in the sonographic evaluation of mineralization using long bones, and that other bones, such as the skull, may be better suited for this assessment (Figures 19-19 and 19-20). The degree of long-bone curvature should also be examined. At present, there is no objective means of assessing this sign, and experience is the only tool to assist the operator in discerning the boundary between normality and abnormality (Figures 19-21 and 19-22). Camptomelia (excessive bowing) is a characteristic of certain disorders (eg, campomelic dysplasia). The possibility of fractures should also be considered, as they can be detected in some conditions (eg, osteogenesis imperfecta) (Figure 19-23). Fractures may be extremely subtle or may lead to angulation and separation of the segments of the affected bone (Figures 19-24 and 19-25). The femur length to abdominal circumference ratio is helpful to predict fetal outcome when a possible skeletal dysplasia is suspected. A ratio below 0.16 has resulted in a lethal outcome. A ratio above 0.16 has predicted a nonlethal outcome.²¹⁴

When performing prenatal diagnosis, it is extremely important to try and distinguish between those cases in which a primary bone dysplasia is actually present versus cases of aneuploidy in which the long bones may be shortened as well. The same applies to cases in which the findings of short limbs are secondary to intrauterine growth restriction, which can mimic skeletal dysplasia on ultrasound. This can be especially difficult when short limbs are detected in the third trimester, or when gestational age is unclear.

Evaluation of Thoracic Dimensions

Several skeletal dysplasias are associated with a hypoplastic thorax (Figures 19-26, 19-27, and 19-28). Such a finding is extremely significant because chest restriction leads to pulmonary hypoplasia, a frequent cause of death in these conditions. In certain cases, despite the fact that the specific type of dysplasia may not be known, the presence of marked thoracic involvement and pulmonary hypoplasia will allow counseling regarding prognosis. The appropriateness of thoracic dimensions can be assessed by measuring the thoracic circumference at the level of the 4-chamber view of the heart. The thoracic circumference can be measured or calculated by using the following formula:

The thoracic length is measured from the boundary between the neck and the chest to the diaphragm. Tables 19-10 and 19-11 show nomograms used to evaluate the thoracic dimensions in fetuses with a known gestational age. However, when gestational age is uncertain, age-independent ratios can be used. The thoracic-to-abdominal circumference ratio (normal value: 0.77 to 1.01) and the thoracic-to-head circumference ratio (normal value: 0.56 to 1.04) permit evaluation of the transverse thoracic dimensions.²¹⁵

The rib cage perimeter (RCP) and thoracic circumference (TC) were measured in a prospective cross-sectional study of 88 patients with normal pregnancy, and 8 cases known to have skeletal dysplasias.²¹⁶ The RCP-to-TC ratio is independent of gestational age (mean = 0.67 ± 0.04). In 5 cases of skeletal dysplasia, this ratio was decreased, and in 1 case it was increased.

Evaluation of thoracic dimensions is a critical part of the workup, because the cause of death in most lethal skeletal dysplasias is pulmonary hypoplasia secondary to an underdeveloped rib cage (Figures 19-29, 19-30, 19-31, and 19-32). Table 19-12 lists skeletal dysplasias associated with altered thoracic dimensions.

Evaluation of Hands and Feet

The hands and feet should be examined to exclude polydactyly (Figures 19-5 and 19-33), brachydactyly (short digits) (Figure 19-34), syndactyly, and extreme postural deformities such as those seen in DTD. Table 19-13 shows a nomogram of the fetal foot size throughout gestation. Table 19-14 lists disorders associated with hand and foot deformities. Disproportion between hands and feet and the other parts of the extremity may also be a sign of a skeletal dysplasia. Figure 19-35 shows the relationship between femur length and foot length. The ratio of femur length to foot length is nearly constant from 14 to 40 weeks, and the mean is 0.99 (SD = 0.06). A ratio below 0.87 should be considered abnormal.²¹⁷ Although fetuses with skeletal dysplasias have been reported to have abnormally low ratios, more evidence is required to test the diagnostic value of this method.²¹⁸ It is expected that a small proportion of unaffected fetuses may have an abnormal ratio. As in the case of other limb biometric parameters, however, large deviations from the lower limits of normal are likely to be significant.

Evaluation of the Fetal Cranium

Several skeletal dysplasias are associated with defects of membranous ossification, which affects skull bones. Orbits should be measured to exclude hypertelorism (Figure 19-36).^219,220 Other findings that should be noted or ruled out are micrognathia,²²¹ short upper lip, abnormally shaped ear,²²² frontal bossing (Figure 19-37), and cloverleaf skull deformity (Figures 19-38 and 19-39). Table 19-15 presents abnormalities of the skull and face associated with different skeletal dysplasias.

Evaluation of the Fetal Face

Sonographic examination of fetal facial features is of major importance in the assessment and diagnosis of skeletal dysplasias because many of these disorders are associated with typical facial abnormalities.²²³ Sonographic evaluation of the fetal face is easily performed in a high percentage of patients from 16 to 20 weeks of gestation onward. The single, most reliable view in detecting facial abnormalities is the sagittal view. This view permits determination of midface hypoplasia, which occurs in several skeletal dysplasias, such as thanatophoric dysplasia, achondroplasia, campomelic dysplasia, osteogenesis imperfecta type I, and spondyloepiphyseal dysplasia congenita.²²⁴

In median clefting, the central portion of the upper lip is absent, and on the midline sagittal view, no upper lip will be seen. In bilateral cleft lip, the midline view will show a variable appearance, depending on the amount of residual premaxillary tissue present in the midline. In unilateral cleft lip (Figure 19-40), the midline sagittal scan may be relatively normal; however, the parasagittal view will demonstrate the cleft. This fact emphasizes the importance of also scanning the fetal face in the coronal plane, which can confirm a suspected abnormality or provide additional information.

Cleft palate occurs in 66% of patients with cleft lip; however, cleft palate is more difficult to diagnose with ultrasound because of shadowing of facial bones. Three-dimensional ultrasound has been shown to be superior to 2D imaging for the prenatal diagnosis of cleft lip and palate, and therefore may be utilized as well.^{225, 226, and 227} Micrognathia is also frequently observed in cases of skeletal dysyplasia (Table 19-16).^{228, 229, 230, and 231} However, this is often diagnosed subjectively. To provide a more objective means to diagnose micrognathia, Paladini et al proposed the jaw index, which is computed as the ratio between the anteroposterior mandibular diameter and the BPD.²³² In studying a population of 198 malformed fetuses, 11 had micrognathia at birth or autopsy. A jaw index of less than 23 was able to correctly identify all cases of micrognathia, with a false-positive rate of only 2%. Recently, Rotten et al²³³ proposed two parameters to differentiate between micrognathia and retrognathia: the inferior facial angle (Figure 19-41) and the mandible width/maxilla width ratio (Figure 19-42A and B). The inferior facial angle is used to diagnose retrognathia. It is defined as the angle between 2 lines traced on a sagittal profile view of the fetal face: (1) reference line (orthogonal to the vertical part of the forehead, at the level of the synostosis of the nasal bones); and (2) profile line (joining the tip of the mentum to the anterior border of the more protruding lip). In studying a population of fetuses at high risk for facial anomalies, an inferior facial angle of less than 50 degrees identified retrognathia with a sensitivity of 100% and a specificity of 98.9%. The mandible width/maxilla width ratio is computed using transverse sections of the mandible and maxilla, with the actual measurements performed 10 mm posteriorly to the anterior osseous border. A ratio less than 0.785 (between 18 and 28 weeks' gestation) correctly identified micrognathia in 3 cases of Treacher Collins syndrome. In addition, nomograms of mandibular length throughout gestation are available (Table 19-17).²³⁴

Intraorbital and interorbital diameters should also be measured, because hypertelorism may occur in cases of skeletal dysplasia (see Figure 19-36; Table 19-18).

Evaluation of the Fetal Spine

Sonographic assessment of the fetal spine is a component in the examination of a fetus with suspected skeletal dysplasia. The following parameters should be assessed.

Vertebral Bodies

Fetal vertebral bodies are composed of three ossification centers representing the vertebral body and two laminae.^{235, 236, and 237} Abnormalities of the ossification center of the fetal vertebral body may result in bony defects, such as hemivertebrae (see Figure 19-13), butterfly vertebrae, or block vertebrae causing congenital scoliosis (see Figure 19-11). A study of the associated anomalies in 27 cases of prenatally detected hemivertebrae noted that, while 11 fetuses had no other abnormal findings, 16 fetuses had additional associated anomalies.²³⁶ These included cardiac, gastrointestinal, renal, facial, extremity, and cranial anomalies. Seven fetuses had bilateral renal agenesis (Potter syndrome). Only 5 of the fetuses with additional anomalies survived. Rib defects are often associated with thoracic vertebral body anomalies (Figure 19-43).

Platyspondyly denotes a condition characterized by decrease of the distance between the upper and lower plates of the vertebral bodies. It may be diagnosed with current high-resolution sonography (see Figure 19-7; Figure 19-44).²³⁸ An objective assessment may be performed by calculating a ratio between measurements of the vertebral interspace to vertebral body height (see Figure 19-44).²³⁹

Clefting of the vertebrae may be complete or incomplete, coronal or sagittal.²⁴⁰ Coronal vertebral clefts are a result of lack of fusion between the anterior and posterior primary ossification centers after 16 weeks of gestation, and can be observed by sonography in utero.^241,242 Sagittal clefts in the vertebral bodies are believed to represent a localized splitting of the notochord due to adhesions between the ectoderm and endoderm during the embryonic period. These clefts range from a single cleft to multiple clefts. The presence of vertebral clefting in skeletal dysplasias was assessed by searching the database at the International Skeletal Dysplasia Registry.²⁴² Coronal and sagittal clefts were present in 40 different conditions. Coronal clefts were more common than sagittal clefts, and were located mainly in the thoracolumbar region. Clefts were most frequently observed in atelosteogenesis (88%), followed by chondroplasia punctata (79%), dyssegmental dysplasia (73%), Kniest dysplasia (63%), and short rib–polydactyly syndrome (53%).²⁴²

Spinal Curvature

The most common osseous anomaly causing scoliosis is unilateral unsegmented bar with contralateral hemivertebrae.^{236,243, 244, 245, 246, and 247} Spinal dysraphism may occur with congenital scoliosis, and this possibility should be examined carefully. An apparent etiologic relationship exists between neural tube defects and other vertebral anomalies. Siblings of infants with congenital scoliosis have a 4% risk of neural tube defects.²⁴⁷ This increased risk is present in siblings of children with a single hemivertebrae, in addition to multiple vertebral anomalies (with or without neural arch defects). The differential diagnosis of fetal scoliosis includes neural tube defects, large abdominal wall defects, amniotic band syndrome, caudal regression, and hemivertebrae. Nonossification of the lumbar vertebral bodies has been detected in achondrogenesis and other diseases.^{248, 249, and 250}

Evaluation of the Internal Organs

A detailed examination of the organs of the central nervous, cardiovascular, gastrointestinal, and genitourinary systems should be performed in all fetuses with skeletal anomalies. Some syndromes present with specific abnormalities of the internal organs, therefore assisting in the differential diagnoses of these entities. For example, congenital heart disease is a prominent feature of Ellis-van Creveld syndrome and Holt-Oram syndrome.

Evaluation of the Neonate

Despite all efforts to establish an accurate prenatal diagnosis, a careful examination of the newborn is always required.¹⁹⁴ It is important that the evaluation include a detailed physical examination performed by a geneticist, or an individual with experience in the field of skeletal dysplasias and radiograms of the skeleton. Birth length, weight, and head circumference should be measured. This will be especially important in situations where physicians have not appreciated growth failure until early childhood when, in fact, it has been present from the time of birth. Radiographs are necessary, and should include anterior, posterior, lateral, and Towne views of the skull, and anteroposterior views of the spine and extremities and scapula,²⁵¹ with separate films of hands and feet. In the majority of cases, examination of the skeletal radiographs will permit a precise diagnosis because the classification of skeletal dysplasias is largely based on radiographic findings.

In suspected lethal skeletal dysplasias, histologic examination of the chondro-osseous tissue should be performed, because this information may lead to a specific diagnosis. Chromosomal studies should be included as well, because there is a specific group of constitutional bone disorders associated with cytogenetic abnormalities. In rare instances, biochemical studies are helpful (eg, hypophosphatasia). In those cases in which the phenotype suggests a metabolic disorder, such as a mucopolysaccharidosis, DNA restrictions and enzymatic activity assays should be considered. In recent years, major advances have been made in identifying mutations associated with chondrodysplasias. Therefore, molecular diagnosis can be performed for disorders in which a specific genetic mutation has been identified, and it is recommended that DNA be saved in all cases.^{252, 253, and 254}

In chromosomally normal pregnancies, it is known that increased nuchal translucency thickness is associated with an increased risk of major anomalies.^{255, 257, and 258} This also includes skeletal dysplasias, such as achondroplasia, asphyxiating thoracic dysplasia, etc. This was established in a large, multicenter screening project for trisomy 21 (100,000 pregnancies) that used the combination of maternal age and nuchal translucency; in this study, an association between nuchal translucency and a wide range of skeletal dysplasias was found among these patients.²⁵⁵ Several case reports and small series have also suggested that in chromosomally normal fetuses, there may be an association between increased nuchal translucency thickness and skeletal anomalies. Table 19-19 summarizes this data.