The spine in Paget’s disease
- 2.4k Downloads
Paget’s disease (PD) is a chronic metabolically active bone disease, characterized by a disturbance in bone modelling and remodelling due to an increase in osteoblastic and osteoclastic activity. The vertebra is the second most commonly affected site. This article reviews the various spinal pathomechanisms and osseous dynamics involved in producing the varied imaging appearances and their clinical relevance. Advanced imaging of osseous, articular and bone marrow manifestations of PD in all the vertebral components are presented. Pagetic changes often result in clinical symptoms including back pain, spinal stenosis and neural dysfunction. Various pathological complications due to PD involvement result in these clinical symptoms. Recognition of the imaging manifestations of spinal PD and the potential complications that cause the clinical symptoms enables accurate assessment of patients prior to appropriate management.
KeywordsPaget’s disease Vertebra Spinal stenosis Magnetic resonance imaging Computed tomography
Paget’s disease (PD) or osteitis deformans is one of the commonest metabolically active bone diseases, second in prevalence only to osteoporosis, characterised by a disturbance in bone modelling and remodelling because of an increase in osteoblastic and osteoclastic activity. It is more common in people of Anglo Saxon origin and is less common in the Far East, India, Middle East and Africa. The overall prevalence of PD is 3–3.7% and increases with age [1, 2, 3]. By the age of 90 years, the prevalence increases to about 10% . As the aetiology of PD continues to be the subject of debate, it is variously classified as an infection, metabolic disorder and a neoplastic process [1, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]. However, strictly speaking, as the unaffected skeleton is metabolically normal, it does not fulfil the criteria for a true metabolic bone disorder.
The epidemiology of PD shows some significant changing trends in the clinical profilometry. First, recent studies have demonstrated that the incidence and prevalence of PD is gradually declining [16, 17, 18]. Second, there is a general trend towards reduction in the severity of the disease as measured by serum alkaline phosphatase levels [19, 20]. Third, there is a steady increase in the age at presentation by about 4 years per decade  and last, the proportion of patients with monostotic disease is increasing . This in turn means that we are likely to see a higher incidence of monostotic vertebral involvement in the future.
The spine is the second most commonly affected site (53%) , after the pelvis (70%) [21, 22, 23]. The disease is polyostotic in 66% of cases and between 35% [24, 25] and 50% [7, 8] have spinal involvement. Using multimodal advanced imaging, this review article focuses on the pathological processes that underlie the varied spinal manifestations and complications of PD.
In PD the loss of homeostatic control leads to increased osteoblastic and osteoclastic activity and constitutes the background for the main three phases. The initial lytic phase represents a mainly osteoclastic activity, the late osteoblastic phase is characterised by new bone formation, while the intervening mixed phase is seen when there is a combination of osteoblastic and osteoclastic activities [24, 26]. One more phase, “inactive sclerotic phase”, characterised by normal or decreased bone activity, has also been described , when the stimulation of new osteoblast and osteoclast formation ceases. Although the bone is metabolically inactive, it maintains a sclerotic coarsened architecture . In bones with a low trabecular/cortex ratio like the skull, femur and humerus, the early lytic phase is radiologically depicted by a clear leading edge at the interface with normal bone. The lytic phase is usually not detected in bones with a high trabecular/cortex ratio like the vertebra, sacrum and pelvis . Vertebral body involvement at radiological diagnosis is virtually always complete, and therefore the leading edge present in the other affected bones is not seen in the vertebra . These phases can be evident in the same patient and at the same time in different bones including the vertebral column. Although progression of disease occurs within an affected bone, the sudden appearance of bone involvement at new skeletal sites years after the initial diagnosis is uncommon [30, 31].
Osseous changes (vertebral body)
The radiological appearance of vertebral body expansion is characterised on radiographs by an increase in the antero-posterior and lateral vertebral dimensions. However, the height of the vertebra is unchanged. The bony vertebral end-plates are subchondral condensations of trabecular bone and do not represent true bony cortex, i.e. there is no periosteum/endosteum interface at the caudal/cranial aspect of a normal vertebra. The sagittal and lateral enlargement is due to one of the previously described pathomechanisms in the corticated portions of the vertebral body. The pathomechanisms are best appreciated on CT (Fig. 2), where the periosteal contour and the endosteal interface due to apposition and absorption are easily seen. Whatever the mechanism causing vertebral body remodelling, bone expansion is a common denominator in PD.
Progression of the sclerotic phase in the spine leads to “ivory vertebra”, due to an increase in the density of the vertebral body, denser than the normal vertebral bodies. There is no change in the density and size of the adjacent intervertebral disc [40, 41]. Differential diagnoses of “ivory vertebra” include metastases, osteosarcoma, carcinoid and Hodgkin’s lymphoma [29, 36]. The increased size of the vertebral body is a useful clue to the underlying diagnosis of PD. Biopsy may be necessary in some cases when there is no expansion (Fig. 5).
In the exceptionally rare cases of vertebrae presenting in the lytic phase, there is marked osteopenia of the vertebra on radiographs, giving a “ghost vertebra” appearance, because of the almost complete involvement of the vertebral body by the osteolytic process . The lytic phase of PD involving the axis  and a rare pathological collapse of a purely lytic L5 vertebra  have been previously reported. It can be difficult to differentiate the lytic phase of PD from other causes of osteolysis and collapse. By demonstrating the cortical thickening, trabecular hypertrophy and vertebral expansion, CT confirms the lytic phase in the vertebral components due to the higher contrast resolution, the feasibility of “bone window” settings and multiplanar reconstructions.
The trabecular hypertrophy and cortical thickening caused by the osseous involvement by PD results in a variable degree of low signal on both T1- and T2-weighted images. In the severe cases of vertebral sclerosis (ivory vertebra appearance) the whole vertebra demonstrates a diffuse low signal on both T1- and T2-weighted sequences. More commonly however, the signal characteristics are heterogeneous on both T1- and T2-weighted sequences due to changes in the intervening marrow space. PD of the vertebra is easily missed or misinterpreted on MR images in the early stages. This is especially true when there is an absence of classic changes including vertebral enlargement and cortical thickening. This is mainly because PD is a disorder of bone and the vertebral marrow is only secondarily affected. MRI can show discrete areas of marrow signal alteration involving vertebral bodies with low non-specific signal on T1-weighted images and high signal on T2-weighted sequences. The marrow changes due to vertebral involvement by PD are described in detail later.
Osseous changes (posterior elements)
Neural arch involvement can be difficult to evaluate on radiographs alone. The best modality for neural arch assessment is CT with reconstructions.
When periosteal and endosteal apposition occur, a markedly sclerotic cortex can be seen [5, 8]. However, when the basic mechanisms are periosteal apposition and endosteal absorption, CT demonstrates an enlarged marrow space of the neural arch delineated by a thin sclerotic line due to the periosteal apposition (Fig. 4). Both processes result in a decrease in the spinal canal diameter.
Bone scintigraphy using 99-Technetium substrates and SPECT can aid the diagnosis of PD, showing a typical distribution of the disease. Scintigraphy demonstrates increased uptake in all the affected vertebral components [46, 47] because of its sensitivity to osteoblastic activity and can be very useful in showing activity in the posterior neural arch. As a whole body imaging modality, it also provides information about the overall distribution of the disease. Moreover, the fourth sclerotic inactive phase described by Milgram is diagnosed essentially on a bone scan where radiographically affected bone does not demonstrate increased activity due to the metabolic inactivity of the lesion . However, increased activity on bone scintigraphy is usually non-specific and requires comparison with radiographs/CT. Various patterns of uptake are described on bone scintigraphy in PD [21, 46, 48, 49], all of which are non-specific.
Imaging techniques such as PET and PET-CT, can contribute to the differential diagnosis of PD from other disorders in the spine, especially when non-specific PD changes are incidentally discovered, and to assess disease activity after treatment.
In general 18F-fluoride positron emission tomography (PET) allows visualisation of the regional skeletal metabolic activity on the basis of the fluoride uptake on the bone surface when new bone formation and blood flow are greater . The advantages of PET-CT on conventional nuclear bone scintigraphy include superior spatial resolution and more accurate quantification of bone activity. In PD, a general increase in global skeletal blood flow, measured by fluoride deposition, is seen associated with a regional increased uptake in the affected bone. Increased activity in the pagetic bone is usually recorded almost immediately after the injection and it remains high throughout the duration of the study. However, this finding is not specific to pagetic bony changes, giving rise to some false-positives. For this reason, the combination of non-specific activity on the PET image, the findings on the CT scan (PET-CT) and the correlation with other radiological examinations increases specificity .
Occasionally pagetic involvement of the spine can also produce extra-osseous extension resulting in ossification of the anterior/posterior longitudinal ligaments and ligamenta flava. Lateral radiographs and CT sagittal reconstruction demonstrate the ossified anterior and posterior longitudinal ligaments. The ligamenta flava ossification are seen on the AP view as radio-opacities adjacent to the facet joints. Para-vertebral soft tissue swelling seen radiographically or by CT can be due to extra-medullary haematopoiesis. When PD of spine results in extra-medullary haematopoiesis , there is extra-osseous soft tissue with signal characteristics similar to the adjacent vertebral marrow. CT and MRI demonstrate the soft tissue mass adjacent to the vertebral body (see Pathological complications) in communication with the bone marrow within the vertebra .
Whenever vertebral bone marrow signal changes are seen on MRI, PD has to be considered in the differential diagnosis because of its non-specific and varied appearance in the bone marrow, and subtle bony alterations suggestive of PD have to be sought. Paget’s disease can, however, also coexist with other disorders.
Pagetic bone is structurally weak. Despite this, the involved vertebral body and posterior neural arch still have to fulfil the biomechanical role they were designed for. The disease will alter the normal metabolic function in the bone while biomechanical forces in turn modify further the metabolic and structural changes that arise. Various clinical complications can occur secondary to vertebral PD resulting in back pain, spinal stenosis and neural dysfunction. Whilst each of these clinical symptoms can be caused by a variety of pagetic complications, it has to be stressed that PD and its complications can be entirely asymptomatic. Moreover, the symptoms could be due to other coexisting spinal disorders.
Causes of back pain in Paget’s disease
Intervertebral disc disease
Overt fractures of vertebrae, sacrum
Causes of neural dysfunction in Paget’s disease
Posterior expansion of the vertebral body
Posterior neural arch expansion
Facet joint arthritis/overgrowth
Extra-osseous involvement—PD, haematopoiesis, “pseudosarcoma”, malignancy
“Arterial steal” syndrome
Para-vertebral swelling can be seen as an indirect sign of the vertebral fracture in the acute phase. The fracture line itself is best seen on CT with sagittal reconstructions. Vertebral fracture can also contribute to the pain related to spinal stenosis by two mechanisms: posterior vertebral body wall bulging and retropulsion (Fig. 16) of a bone fragment into the spinal canal [8, 68], and cord compression. This can be recognised on radiographs in the lateral view and is associated with a decrease in the spinal canal size, but CT demonstrates this better.
The fractured vertebra in PD sometimes appears osteopenic on radiographs. This could be due to a fracture occurring in the lytic phase of PD. However, as described earlier, it is unusual to see the lytic phase of PD in the spine. The osteopenia is most likely due to fracture-induced osteolysis. MRI can also contribute to the diagnosis of a compression fracture, even if its findings are not specific. The fracture line is seen as a linear or curvilinear low signal on T1-weighted images and high signal on T2-weighted and STIR images. Oedema surrounding the fracture can be an indirect sign. It is seen as high signal on inversion recovery and T2-weighted images and low signal on T1-weighted images. When the fracture is not recent, the signal on the MRI scan will be low on both T1- and T2-weighted images because of the sclerosis of the bone. In general, diffusion weighted imaging can help to differentiate osteoporotic from malignant vertebral body collapse . It is not yet known if benign fracture in PD can be mistaken for a malignant compression on diffusion weighted imaging given the associated underlying marrow changes prior to the occurrence of the fracture.
Facet joint arthropathy
Facet joint involvement can cause back pain, lateral spinal stenosis and neural dysfunction. Facet involvement can lead to facet joint arthropathy due to two mechanisms. One is the direct extension of PD into the articular cartilage similar to that seen in large synovial joints and the intervertebral disc , the other is the pagetic involvement of the facets leading to expansion of the facet and resultant incongruity with the apposing uninvolved facet (Fig. 9). Facet joint arthropathy was noted in 17 out of 21 symptomatic patients with back pain in one study . This gives rise to mechanical stress and early degenerative changes in the facet joint. This is one of the most important factors in back pain and also contributes to spinal stenosis. The spinal canal stenosis associated with facet joint arthropathy is typically lateral and can result in nerve root compression . Patients with severe arthropathy often have symptoms of spinal stenosis, but they can be asymptomatic [24, 25]. Compared with radiography, CT gives a clearer depiction of facet joint changes (Fig. 9), showing the loss of articular space, the destruction of cartilage associated with subchondral erosion and articular incongruity (Fig. 9) [5, 26, 70]. Facet joint arthropathy may eventually lead to ankylosis of the articular space . MR imaging can underestimate the bony changes associated with the facet involvement.
Spondylolysis and spondylolisthesis
Spondylolisthesis in PD can be due to either spondylolysis or facet joint degeneration (Fig. 17). It is important to differentiate between the two entities. The degenerative spondylolisthesis results in spinal stenosis and neural dysfunction. The lateral radiograph is quite useful for detecting spondylolisthesis and to establish its grade and severity. However, this can be underestimated if the vertebral alignment is not assessed carefully. An enlarged vertebra due to PD may slip anteriorly over a normal vertebra, but the posterior vertebral body alignment may still be preserved due to the posterior expansion. The assessment of the posterior neural arch alignment, however, will reveal the subtle slip in these cases. The standard lateral radiograph can be complemented by dynamic views in flexion and extension, which can reveal the presence of spondylolisthesis under stress. Spondylolisthesis, and particularly spondylolysis, can be confirmed on CT with sagittal reconstruction images.
Intervertebral disc involvement
Neoplastic transformation in pagetic bone is very rare in the spine (0.7%) [5, 8, 24, 72] and represents only 7%  of all sarcomatous degeneration in PD. It can present with back pain. Benign and malignant bone neoplasms can complicate Paget’s disease. The majority of malignant transformations are osteosarcomas. They seem to be related to one or two genes on the chromosome 18q [5, 74, 75].
Causes of paravertebral swelling in Paget’s disease
Paget’s sarcoma/other tumours
Recognition of the imaging manifestations of spinal PD and the potential clinical complications enables accurate assessment of patients prior to appropriate management. This knowledge should allow subtle PD to be identified on imaging when this is not suspected. This is especially relevant to MRI, as it has become the imaging modality of choice for investigating the spine. Patients presenting with back pain and spinal stenosis have to be assessed for pagetic complications before attributing the symptomatology to the disease itself.
- 27.Milgram JW. Radiographical and pathological assessment of the activity of Paget’s disease of bone. Clin Orthop Relat Res 1977; 127: 43–54.Google Scholar
- 29.Epstein BS. The spine: a radiologic text and atlas, 4th edn. Lea & Febiger, Philadelphia, 1976.Google Scholar
- 40.Ochsner HC, Moser RH. Ivory vertebra. Am J Roentgenol 1933; 29: 635–637.Google Scholar
- 41.Souques A, Lafourcade, Terris. “Vertebra d’ivoire” dans un cas de cancer metastatique de la colonne vertebrale. Rev Neurol 1925; 32: 3–10.Google Scholar