Osteodysplasias are well‐defined abnormalities of bone growth, typically genetic in origin. They may be systematic, affecting one aspect of skeletal development, typically in proportion to growth potential of each site, or else “aleatoric,” affecting sites of an aspect of bone development in a chance distribution, with a higher likelihood of occurrence at sites of greater normal growth. [I have borrowed the term aleatoric from modern music composition in which chance plays a role ( 1)] The aspect of growth may be slowed, disordered, or accelerated. The aspect involved is most often exclusively enchondral in nature or membranous in nature, with only secondary changes in the other. Most, but not all, dysplasias result in short stature (dwarfism). Kindred disorders include dysostoses, in which individual aspects of growth are affected, and sequences, in which an abnormal aspect of development causes a chain of consequent effects on bone (and other tissue) development. As more genetic markers for dysplasias are determined, more conditions are being grouped into “families” of dysplasias of like genetic origin.
Disproportionate short stature is found in many systemic dysplasias; limb length discrepancy, with limp, is frequent in aleatoric dysplasias. Fractures (and blue sclerae) are a hallmark of osteogenesis imperfecta; teeth tend to show dentinogenesis imperfecta crowding and irregularity as well. Tall stature is seen in the dolichostenomelias, notably Marfan syndrome and homocystinuria. With regard to characterization of short limbs in dysplasias, the term rhizomelic refers to shortening of limbs most pronounced proximally, such as achondroplasia; mesomelic dysplasias have most of the relative shortening in the intermediate segments (radius/ulna and tibia/fibula); and acromelic refers to the greatest shortening distally. Other rather specific findings include hitchhiker thumb in diastrophic dysplasia; thumb extension beyond the fist in Marfan syndrome; early‐in‐life shortening, especially of the limbs and later shortening of the trunk in metatropic dysplasia; gargoyle‐like facies in the mucopolysaccharidoses; and occasionally the ability to draw the shoulders together in front of the chest in cleidocranial dysplasia/dysostosis.
Skeletal surveys for osteochondrodysplasia or dysostosis need to be customized for the suspected diagnoses; once certain findings appear, other views may be required. For example, whenever platyspondyly (flatter than normal vertebral bodies) is observed, one needs to check the dens and C1 region in detail, with careful flexion and extension views as necessary, to evaluate for associated subluxability between C1 and C2. The small tubular bones of the hand and feet should be included in any survey because they may show diagnostic features key to the diagnosis. Tall patients with dolichostenomelia (such as Marfan) need good lateral chest or thorax images to evaluate for pectus excavatum or carinatum. In aleatoric disorders such as multiple exostosis or enchondromatosis, two orthogonal views of each part are needed for complete evaluation of the lesions, even if not necessary for diagnosis.
In utero imaging for dysplasia or dysostosis begins with high‐detail ultrasound. Questions that arise may then be further investigated with magnetic resonance imaging (MRI). If the answer is still not given, but is important, selected radiographs of the mother's abdomen may be considered, under close supervision by a specialized pediatric radiologist. If possible, for radiation protection considerations, radiographs of one twin should be avoided if the other is considered unaffected. Abnormal bone length, shape, or a positive family history of dysplasia or dysostosis should initiate the careful prenatal evaluation for specific entities.
MRI can give insight into the nature and quantity of cartilage in several dysplasias, including Kniest disease ( 2). MRI of the spinal cord is prudent before planned spinal surgery in children with any severely abnormal spinal curvature. Three‐dimensional reconstruction of abnormal joints assists in surgical planning for specific procedures.
Accurate measurement of long bone lengths requires perpendicular X‐ray beam images at each end of the bone with a radiopaque ruler alongside that is not shifted between the exposures. Alternately, measurements can be made from a computed tomography (CT) scout view.
Long bones for dysplasia evaluation need to be parallel to the film or screen, lest foreshortening in space be mistaken for dysplastic growth foreshortening through time.
Multiple cartilaginous exostoses each have a cartilage cap with a growth plate (I call it “paraphysis”) that will have uptake on bone scan similar to physes. Fractures in such conditions as osteogenesis imperfecta, Ollier enchondromatosis, and osteopetrosis will be hot on bone scan. The periosteal reaction causing the thick cortices of Engelmann disease will be hot on bone scan in the active stage, as it will be in other membranous bone overactivity dysplasias or dysostoses, such as van Buchem disease. Osteopetrosis usually gives a superscan of high activity of all involved bone segments. Uptake is also increased in active fibrous dysplasia.
Only a few dysplasias and dysostoses will be described here.
Systematic dysplasias are symmetric side‐to‐side and generally change one rule or rate of growth, with the sites of greatest normal growth being affected the most. The prototype is achondroplasia with rhizomelic shortening; lumbar pedicles abnormally close side‐to‐side with the distance decreasing downward, greater than normal concavity of medial and lateral margins of metaphysis (Fig. 1), frontal bossing of the skull, petrous ridges closer to each other than normally, frequent cartilage rests in metaphyses, smooth but delayed growth centers with epiphyseal shape following the shape of abnormally concave physes, proximal ends of the femurs in infancy showing the pattern of an ice cream scoop on end, horizontal sacrum, shorter than normal metaphyseal collar, and trident (Vulcan salute) hands, among the many imaging findings. Hypochondroplasia, an allelic (due to the same gene locus) disorder, is less severe; thanatophoric dysplasia, also allelic, is more severe (demonstrating femurs resembling European telephone receivers and quite pronounced platyspondyly). Achondroplasia and its family results from slowed enchondral growth, whereas Marfan syndrome and homocystinuria have dolichostenomelic (predominant proximal segment) lengthening due to accelerated enchondral growth. The medial and lateral margins of long bone metaphyses are less concave than normal. Children with Marfan syndrome show pectus carinatum or excavatum, tortuous aortic arch, and wide lumbar canal; homocystinuria shows truncal osteoporosis.
In metatropic dysplasia, the metaphyseal collar bone bark seems to lack its usual ability to restrict too rapid transverse growth of physis and metaphysis, so that bones resemble dumbbells with unusually broad metaphyses. The megaepiphyseal dysplasia Kniest disease shows coronal cleft vertebral bodies and delayed ossification of the (large) centers for the femoral head. Severe cervical kyphosis is seen (when lateral images are obtained) in diastrophic dysplasia and camptomelic dysplasia and is one of the cervical vertebral anomalies seen in Larsen syndrome. A small mandible with an abnormally concave undersurface is the key finding of Pierre‐Robin sequence and is seen in camptomelic dysplasia, cerebrocostomandibular syndrome, and Seckel syndrome. Progressive pseudorheumatoid dysplasia has joint region changes closely resembling rheumatoid arthritis, but also has both early and unusually large os trigonum centers behind the talus, and bullet‐shaped or Scheuermann‐like (irregular endplates) thoracolumbar vertebral bodies unlike rheumatoid arthritis. Every normal child, incidentally, has at least one small os trigonum center that appears near the end of the first decade of life and fuses with the talus in about a year, appearing somewhat earlier in girls than in boys. Then, in the second decade, another os trigonum center appears in some 10–20% of children, which may or may not also fuse to the talus to remain as a posteriorly protruding process.
Postaxial polydactyly and mesomelic short limbs are characteristic in Ellis–van Creveld syndrome. An interesting accompaniment to hand polydactyly is the ham‐shaped hamate (a wider than normal hamate ossification in the form of a cured ham).
Various characteristic vertebral body shapes on lateral images occur in the several spondyloepiphyseal and spondylometaphyseal dysplasias, as well as mucopolysaccharidosis dysostosis multiplex conditions. Twisted ribs and long bones are the prime feature of Melnick–Needles osteodysplasty.
On in utero imaging, achondrogenesis, the second most frequent lethal skeletal dysplasia, is characterized by a lack of ossified vertebral bodies. Very short ribs, such as in short rib polydactyly syndromes, also predict neonatal demise, as do features of thanatophoric dysplasia (the best‐known lethal dysplasia). The shorter the ribs in Jeune syndrome, the less likely the neonatal survival.