Hemangiomas

Hemangiomas are benign vascular tumors that occur in more than 10% of adults and are commonly detected as an incidental finding on imaging studies performed for unrelated indications. They are the most common primary bone tumor in adults, occur most commonly in the thoracic spine, and are usually solitary but can be multiple in approximately 30% of cases.^1,2,21 They typically arise in the vertebral body but may involve the posterior elements. The CT appearance of a low attenuation lesion with coarse trabeculae throughout (giving a “polka-dot” appearance in cross-section) is diagnostic.²² MRI demonstrates the fatty stroma, which is bright on T1WI and iso-intense to hyperintense to marrow on T2WI, with avid enhancement after administration of gadolinium.²³ Bone scan is typically normal.²⁴ An aggressive subtype of hemangioma is recognized that tends to be associated more commonly with epidural extension and pathological fracture. These lesions are often iso-intense to hypointense to marrow on T1WI and can be impossible to distinguish from a malignant lesion, such as a metastasis, on imaging (Table 9-3 and Figs. 9-3 to 9-6).^25,26

Table 9-3 Features of Hemangiomas

STIR, Short tau inversion recovery.

Fig. 9-3 Typical benign hemangiomas found incidentally in a 70-year-old woman imaged for back pain. A, Sagittal T1WI of the lumbar spine demonstrates round, relatively well-circumscribed hyperintense lesions within the T12 and L2 vertebral bodies. Mild endplate degenerative changes are present at L5–S1. The remainder of the visualized bone marrow is normal for a patient of this age. B, On the sagittal T2-weighted FSE image the two lesions remain hyperintense to the normal bone marrow. C, The two lesions are dark on the STIR image, blending in with normal bone marrow. D, Axial T2WI at the T12 level demonstrates the hyperintense well-circumscribed hemangioma within the left side of the vertebral body.

Fig. 9-4 Typical benign hemangiomas may sometimes have high signal on STIR images. A, Sagittal T1WI demonstrates a well-circumscribed hyperintense lesion within the L2 vertebral body consistent with a hemangioma. B, Sagittal T2WI demonstrates high signal within this lesion confirming the diagnosis of hemangioma. C, Sagittal STIR image. Unlike the previous example in Figure 9-3, high signal is seen within the L2 hemangioma on STIR. This does not change the diagnosis of hemangioma.

Fig. 9-5 Legend continued on page 129.

Aggressive T9 hemangioma with epidural and paravertebral extension and mild wedging of the vertebral body inferior endplate. A, Midline sagittal T1WI of the thoracic spine demonstrates a heterogeneous hyper-intense lesion involving almost the entire T9 vertebral body. B, On the sagittal T2WI, the lesion is heterogeneous but predominantly hyper-intense. Mild posterior bulging of the T9 vertebral body cortex is present with effacement of the CSF space anterior to the lower thoracic cord. C, Off-midline sagittal post-gadolinium T1-weighted fat saturated image demonstrates the heterogeneously enhancing hemangioma with an epidural component encroaching on the neural foramina above and below T9 and extension into the posterior elements. Mild central wedging of the vertebral body is noted with buckling of the inferior vertebral body endplate. D, Axial post-gadolinium T1-weighted fat saturated image at the T9 level. The extent of paravertebral and epidural involvement is appreciated best in the axial plane. Fat saturation nulls signal from normal epidural fat and is important to characterize encroachment on the spinal canal.

Fig. 9-6 Aggressive hemangioma with a large epidural component. A, Off-midline sagittal T1WI demonstrates heterogeneous mildly hyper-intense lesion within the T11 and T12 vertebral bodies with soft tissue encroaching on the T11–12 neural foramen, obliterating the normal perineural fat in this location. B, Midline sagittal T1WI demonstrates the heterogeneous hyper-intense T11 and T12 vertebral body lesions as well as the large associated posterior epidural soft tissue mass compressing the cord at these levels. An epidural fat-cap is seen along the superior margin of the epidural mass, distinguishing it from an intradural lesion. C, Midline T2WI shows both the vertebral body lesions and epidural component to be hyper-intense to normal marrow and soft tissue structures. D, Off-midline post-gadolinium T1-weighted fat saturated image. The epidural component of the aggressive hemangioma, extending into the intervertebral neural foramen, enhances brightly and homogeneously. The vertebral body components enhance heterogeneously and slightly less avidly. E, Midline post-gadolinium T1-weighted fat saturated image. The brightly enhancing epidural component of the lesion stands out against the dark epidural fat. F, Axial post-gadolinium T1-weighted fat saturated image at the T12 level. The degree of cord compression by the enhancing epidural mass is appreciated best in the axial plane. Similarly, involvement of the left paraspinal soft tissues and left costovertebral junction is best appreciated in the axial rather than sagittal plane.

Metastases

Metastatic disease is defined as dissemination or extension of tumor by direct, hematogenous, or lymphatic routes. Metastases are the most common malignancy to affect the spinal column. Bone metastases occur in 50% of all cancer patients, and 40–70% of these lesions are located within the vertebrae.²⁷ In adults, breast, lung, prostate, lymphoma, sarcoma, and kidney account for the majority of primary sites. In children, neuroblastoma and Ewing sarcoma are the most common primary malignancies to metastasize to the spine.^21,28 Metastases typically involve the vertebral body and posterior elements, and compression fractures as well as extension into the epidural space are common features. Most lesions have a lytic appearance on radiographs and CT, but blastic or sclerotic metastases can occur, especially in the setting of prostate carcinoma.^1,2 Bone scan traditionally has been used to screen for metastatic disease to the spine. It has a sensitivity of approximately 95%, but can have false negatives if there is only marrow infiltration without cortical involvement, and is often non-specific. MRI has been shown to be both more sensitive and specific than scintigraphy.³ Radiographs are insensitive to assess for metastatic disease, requiring at least 50–70% bone destruction for detection of lesions.^27,29,30 Classic radiographic signs include an absent or sclerotic pedicle, cortical destruction, and paraspinal soft tissue mass.^29–31 CT is more sensitive than radiographs to detect bony destruction, sclerosis, and paraspinal masses but is less sensitive than MRI and is poor for assessment of cord compression. On MRI, lesions are typically hypointense to normal marrow and intervertebral discs on T1WI, usually hyperintense on T2WI, and demonstrate heterogeneous enhancement.^1,2 In the case of epidural extension, the “draped curtain” sign has been described, whereby there is sparing of the midline because of an intact midline septum that attaches the dura anteriorly to the posterior longitudinal ligament, in contrast to infection, which does not spare the midline.³² When pathologic fracture occurs as a result of an underlying metastasis, restricted diffusion may be helpful to distinguish the metastatic lesion from a benign osteoporotic fracture^14–17 (Table 9-4 and Figs. 9-7 to 9-10).

Table 9-4 Features of Spinal Metastases

Fig. 9-7 75-year-old male with prostate cancer and diffuse metastatic disease throughout the spine. MRI of the lumbar spine was performed to rule out cord compression. A, Sagittal T1WI. Bone marrow signal is heterogeneous and lower than the adjacent intervertebral discs because of diffuse infiltration with metastatic disease. Vertebral body heights are maintained with no evidence of fracture. B, Sagittal T2WI demonstrates heterogeneous marrow signal with areas of high and low signal that both correspond to metastatic deposits. C, Sagittal post-gadolinium T1-weighted fat saturated image. Diffuse heterogeneous enhancement is present throughout the lumbar spine. A small amount of enhancing epidural soft tissue is present, but there is no evidence of cord or nerve root compression. D, Axial post-gadolinium T1-weighted fat saturated image. The extent of enhancing epidural soft tissue is appreciated best in the axial plane.

Fig. 9-8 A 46-year-old man with known history of colon cancer. MRI of the thoracic spine demonstrates solitary T4 metastasis. A, T1WI demonstrates low signal lesion involving the entire T4 vertebral body and bowing the cortex posteriorly with cortical disruption and epidural extension. B, Sagittal post-gadolinium T1-weighted fat saturated image demonstrates heterogeneous enhancement of the lesion. No other lesions are identified. C, Axial post-gadolinium T1-weighted fat saturated image demonstrates diffuse involvement of the vertebral body and cortical disruption posteriorly in the midline with epidural extension.

Fig. 9-9 Legend continued on page 135.

MRI vs. CT visualization of metastatic disease. This patient has diffuse metastatic disease throughout the spine. Spinal decompression and fixation were performed extending from T11 to L2, with associated artifact from the posterior metallic hardware. A, Sagittal CT image demonstrates diffuse mottled appearance of the spine because of numerous lytic lesions. Lesions causing cortical destruction are easy to appreciate. However, those without cortical disruption are subtle and easy to overlook. Note the beam hardening artifact and streak artifact in the lower spine that are a result of metallic hardware. B, Axial CT image demonstrates a focal lytic metastatic lesion within the posterior vertebral body with associated cortical disruption. If cortical disruption had not been present, this lesion would be easy to dismiss as a Schmorl’s node related to degenerative change. C, Sagittal T1-weighted MRI image demonstrates marrow heterogeneity, but diffuse metastatic disease is difficult to confirm in this case without administration of gadolinium. D, Sagittal T2-weighted MRI image. The presence of diffuse metastatic disease throughout the vertebrae cannot be appreciated on this sequence. The FSE technique is helpful in minimizing artifact from the metallic spinal hardware. However, note the presence of prominent foci of low signal from CSF flow artifact within the posterior CSF space. This appearance should not be confused with intradural metastases. E, Sagittal post-gadolinium T1-weighted fat saturated image. Numerous enhancing lesions are evident throughout the vertebral bodies and within the posterior elements at multiple levels. Artifact from the metallic hardware is worst on this sequence and interferes with fat saturation at the adjacent levels.

Fig. 9-10 Patient with non-small cell lung cancer presenting with back pain and FDG PET avid metastatic deposit to right T6 pedicle. A, Sagittal T1WI of the thoracic spine demonstrates a focal low intensity lesion within the right T6 pedicle. No other spinal lesions were seen. B, Sagittal T2-weighted FSE image. The lesion is hypo-intense to normal marrow. C, Sagittal STIR image. The lesion is hypo-intense with a hyper-intense rim. D, Sagittal post-gadolinium T1-weighted fat saturated image. The rim of the lesion enhances avidly, whereas the central portion of the lesion only enhances mildly. E, Axial T2-weighted FSE image at the T6 level. There is no encroachment on the spinal canal. F, Axial post-gadolinium T1-weighted fat saturated image at the T6 level. G, Axial FDG PET image at the T6 level demonstrates increased uptake within the involved T6 pedicle. There is also increased uptake within involved right hilar lymph nodes. H, Axial FDG PET image at a lower thoracic level demonstrates avid FDG uptake within the primary lung lesion at the right lung base. I, Axial CT image at the T6 level. The T6 pedicle lesion is not seen because there is no cortical destruction. The enlarged right hilar lymph nodes are demonstrated on this image.

Pathological vs. Benign Compression Fractures

Vertebral compression fractures in the absence of trauma are a common clinical problem in the elderly population. Although clinical history is helpful, up to one-third of fractures in patients with known primary malignancy are benign, and approximately one-quarter of fractures in apparently osteopenic patients are caused by metastases.³¹ Diagnosis of an underlying lesion is important because it influences clinical staging, treatment planning, and prognosis for the patient. In the chronic setting, the differentiation between pathologic fracture as a result of underlying malignancy and benign osteoporotic fracture is fairly simple and can be made with a high level of certainty.^33,34 Marrow signal of chronic benign fractures is iso-intense to normal bone marrow on all sequences, whereas fractures associated with metastases demonstrate low signal intensity on T1-weighted sequences and high signal intensity on T2-weighted sequences.^33,35 STIR images provide the greatest contrast between normal and abnormal bone marrow.³³ Acute compression fractures, however, may share many of the imaging findings of metastatic lesions, and differentiation is more challenging.^35,36 Features that favor acute benign osteoporotic fractures include retropulsion of a bony fragment, preservation of normal marrow signal intensity, a horizontal band-like pattern of low signal intensity on T1WI and T2WI, and the presence of other compression fractures.^37,38 Features more likely to be seen in metastatic compression fractures include pedicle involvement, an associated focal paraspinal soft tissue mass or epidural mass (particularly one encasing the dural sac), convex posterior cortex, diffuse low signal intensity within the vertebral body on T1WI, and the presence of other metastases.^35,37–30 Some studies have suggested that a pattern of intense or heterogeneous contrast enhancement supports diagnosis of metastatic compression fracture, but this finding is inconsistent in the literature and may not be reliable.^34,35,37,39 There has been increasing interest in the use of diffusion weighted imaging for differentiating benign vertebral fractures from those associated with metastases, with early results showing that reduced diffusion is highly specific for diagnosing an underlying metastatic lesion.^14–17 However, even this technique is not foolproof¹⁸ and either bone biopsy or follow-up imaging is often required (Table 9-5 and Figs. 9-11 to 9-14).

Table 9-5 Imaging Features to Differentiate Benign Fracture from Malignancy

Fig. 9-11 75-year-old female with chronic osteoporotic compression fractures. A, Sagittal T1WI of the lumbar spine demonstrates endplate compression and varying degrees of loss of height of all of the lumbar vertebral bodies. Bone marrow signal is uniform throughout the spine and is high signal relative to the intervertebral discs. B, Sagittal T2WI also demonstrates uniform normal signal throughout the spine as is expected for chronic benign compression fractures. C, Sagittal STIR image shows uniform low signal within the bone marrow with no evidence of bone marrow edema as would be seen with acute fractures or with an underlying lesion. D, Sagittal post-gadolinium T1-weighted fat saturated image shows no abnormal enhancement.

Fig. 9-12 A 66-year-old woman with vertebral metastases and pathologic chronic L2 compression fracture. There is evidence of prior pelvic radiation with postradiation changes seen within L5 and the sacrum. A, Sagittal T1WI of the lumbar spine demonstrates low signal lesions within the T12, L2, and L3 vertebral bodies. There is pathologic fracture through the L2 vertebral body with approximately 30% loss of vertebral body height. The L5 vertebral body and sacrum are uniformly high in signal because of prior radiation treatment to this area. B, Sagittal T2-weighted FSE image. The T12, L2, and L3 lesions have low signal on this sequence. C, Sagittal post-gadolinium T1-weighted fat saturated image. The vertebral body metastases enhance. Chronic benign vertebral body fractures should not have this pattern of enhancement.

Fig. 9-13 Metastatic disease with acute L4 fracture and dural sac encasement. A, T1WI of the lumbar spine demonstrates compression fracture of the L4 vertebral body with greater than 50% loss of vertebral body height. The heterogeneous low signal within the vertebral body could be a result of edema rather than metastatic infiltration. B, T2WI of the lumbar spine demonstrates the acute fracture but does not demonstrate specific features of benign vs. pathological fracture. C, Sagittal STIR image. On this sequence, there is high signal within the L4 vertebral body that could be caused by edema from the acute fracture. However, additional high signal lesions are evident within the L3, L5, and S1 vertebrae, consistent with metastases that could not be seen on the conventional T1- and T2-weighted sequences. D, Sagittal post-gadolinium T1-weighted fat saturated image. The hyper-intense lesions seen on the STIR sequence enhance as expected for metastatic lesions. The acute fracture enhances, which is a non-specific finding and could be seen with both benign and pathological fractures. E, Axial post-gadolinium T1-weighted fat saturated image at the L4 level demonstrates the enhancing fractured vertebral body. There is also a small amount of epidural enhancing material encasing the thecal sac, which increases the likelihood that this represents a fracture caused by underlying malignancy.

Fig. 9-14 Legend continued on page 142.

24-year-old female with acute traumatic compression fractures. A, Sagittal T1WI. There is subtle central wedging of the T12 vertebral body and anterior wedging of the L1 vertebral body with minimal, if any, loss of vertebral body height. Subtle linear low signal is present within the superior T12 and L1 vertebral bodies along the fracture lines. Bone marrow signal is otherwise uniform throughout the lumbar spine and normal for a young patient. Note that bone marrow signal on the T1-weighted sequence is slightly lower than that seen in examples of older patients, but is still greater than the adjacent intervertebral discs. B, Sagittal T2-weighted sequence. The fractures extending through the T12 and L1 vertebral bodies are more readily appreciated as a result of high signal from associated edema. C, Sagittal STIR. This sequence is most sensitive for demonstrating the high signal caused by edema from the acute fractures. D, Sagittal post-gadolinium T1WI. The acute fractures demonstrate linear enhancement. E, Sagittal CT reconstruction. The T12 and L1 fractures are much more difficult to see than those on the MRI. Subtle anterior cortical buckling is noted.

Multiple Myeloma

Multiple myeloma is a multifocal malignant proliferation of monoclonal plasma cells that occurs most commonly in men older than 60 years. Diagnosis is confirmed by bone biopsy or by demonstrating Bence Jones proteins (free light chains) in urine or monoclonal gammopathy in serum. Although it is the most common primary bone malignancy, multiple myeloma accounts for only 1% of all cancers.³¹ The axial skeleton is more commonly involved than long bones. Vertebral body destruction and fractures are common with spine involvement.⁴⁰ Punched out lytic bone lesions, diffuse osteopenia, fractures, and, rarely, sclerotic lesions are the hallmarks of disease on CT and radiographs.¹ MRI findings reflect a number of different patterns of bone marrow involvement. These patterns include microscopic infiltration with normal MRI appearance in up to 20%, focal lesions, homogeneous diffuse infiltration of the bone marrow (best seen as low signal on T1WI), combined diffuse and focal disease, and a heterogeneous/variegated pattern with interposition of fat islands giving a “salt and pepper” appearance.⁴¹ Marrow involvement or lesions are typically hypointense on T1WI, hyperintense on T2WI and STIR images, and enhance avidly with gadolinium. MRI is helpful for following treatment response, with decreased T2 signal abnormality and decreased enhancement representing good prognostic signs.^42,43 Bone scan has limited sensitivity, detecting bone involvement in 75% of myeloma patients and only demonstrating 10% of lesions. However, technetium (Tc)-99m-sestamibi scintigraphy is both sensitive and specific for diagnosing myeloma lesions and is complementary to bone scan. Fluorodeoxyglucose (FDG) PET is gradually assuming an increasingly important role in following response to treatment⁴⁴ (Table 9-6 and Figs. 9-15 and 9-16).

Table 9-6 Features of Multiple Myeloma