MRI Features of Extramedullary Myeloma

Transcription

1 Musculoskeletal Imaging Original Research Tirumani et al. MRI of Extramedullary Myeloma Musculoskeletal Imaging Original Research Sree Harsha Tirumani 1,2 Atul B. Shinagare 2 Jyothi P. Jagannathan 1,2 Katherine M. Krajewski 1,2 Nikhil C. Munshi 3 Nikhil H. Ramaiya 1,2 Tirumani SH, Shinagare AB, Jagannathan JP, Krajewski KM, Munshi NC, Ramaiya NH Keywords: extramedullary myeloma, MRI, multiple myeloma DOI: /AJR Received March 2, 2013; accepted after revision June 23, Department of Imaging, Dana Farber Cancer Institute, Harvard Medical School, 450 Brookline Ave, Boston, MA Address correspondence to S. H. Tirumani (stirumani@partners.org). 2 Department of Radiology, Brigham and Women s Hospital, Harvard Medical School, Boston, MA. 3 Department of Hematologic Oncology, Dana Farber Cancer Institute, Harvard Medical School, Boston, MA. This article is available for credit. AJR 2014; 202: X/14/ American Roentgen Ray Society MRI Features of Extramedullary Myeloma OBJECTIVE. The purpose of this study was to describe the MRI features of extramedullary myeloma and to evaluate the role of MRI in extramedullary myeloma. MATERIALS AND METHODS. The cases of 28 patients (15 men, 13 women; mean age, years; range, years) with extramedullary myeloma who underwent MRI at one institution from January 2004 through December 2012 were retrospectively identified through an electronic search of an institutional radiology database. Two radiologists reviewed images from 44 MRI examinations in consensus to document the morphologic, signal-intensity, and enhancement characteristics of extramedullary myeloma. Electronic medical records were reviewed to document the indication for MRI and subsequent management of extramedullary myeloma. RESULTS. A total of 72 sites of extramedullary myeloma were noted, most commonly the paraspinal-epidural location (28/72, 39%). Two radiologic patterns were identified: lesions contiguous with bone (n = 44) and lesions noncontiguous with bone (n = 28). Lesions contiguous with bone were larger (p = 0.001; Student t test). Of 28 paraspinal-epidural lesions, 13 compressed the cord. Compared with skeletal muscle, most of the lesions were hypointense to isointense on T1-weighted images (67/72, 93.1%) and isointense to hyperintense on T2-weighted images (62/72, 86.1%). Lesions noncontiguous with bone were more often hypointense on T2-weighted images (8/28 vs 2/44; p = 0.006; Fisher exact test). Neurologic symptoms prompted MRI in most cases (n = 32/44). MRI was helpful in management by radiotherapy and surgery (19/28). CONCLUSION. Extramedullary myeloma can be contiguous or noncontiguous with bone. Lesions contiguous with bone are larger, often occur in a paraspinal or epidural location, and can cause cord compression. Lesions noncontiguous with bone can be T2 hypointense. MRI helps in treatment planning. M ultiple myeloma (MM) is the most common primary bone malignancy, commonly affecting patients years old, men more often than women [1]. Important advances have been made in the diagnosis, staging, and treatment of MM, increasing the overall survival of these patients to more than 10 years from the initial time of diagnosis [2]. With this longevity, there has been a concurrent increase in the unusual manifestations of MM, especially relapses in extramedullary sites, referred to as extramedullary myeloma [3]. Extramedullary myeloma is an unusual presentation of MM, and little is known about its incidence and natural history. Extramedullary myeloma is seen in 7 18% cases of newly diagnosed myeloma and in 6 20% of cases during the disease course [4]. In as many as 45% of patients with extramedullary myeloma, the tumor develops at the time of relapse, particularly in patients treated with allogenic bone marrow transplant, in whom the extramedullary myeloma occurs as an escape phenomenon [3, 5]. The pathophysiologic characteristics of extramedullary myeloma are elusive. Bladé et al. [5] described two mechanisms to explain the development of extramedullary myeloma (Fig. 1). The first mechanism, which is by far the more common, involves contiguous extraskeletal extension of myelomatous masses. The less common second mechanism involves hematogenous dissemination of a subclone of myeloma cells that have decreased expression of cell surface adhesion receptors, allowing bone marrow escape. The most common sites of hematogenous spread are the skin, viscera, AJR:202, April

2 lymph nodes, upper airways, and CNS. Biologically, extramedullary myeloma tends to behave as high-grade lymphoma and is often associated with complex genetic abnormalities on gene expression profiles [3]. The presence of extramedullary myeloma usually indicates a poor prognosis due to decreased overall survival and progression-free survival [3]. There are no established guidelines for the imaging workup of patients with suspected extramedullary myeloma. The International Myeloma Working Group recommends the use of 18 F-FDG PET/CT for all patients with suspected extramedullary myeloma [6]. However, both CT and MRI play key roles in the management of extramedullary myeloma by facilitating evaluation of site-specific symptoms [7]. MRI, because of its higher soft-tissue resolution, can help in better characterization of extramedullary myeloma and better visualization of marrow involvement. The literature on the MRI features of extramedullary myeloma is restricted to small case series and a few isolated case reports. The aim of our research was to evaluate the MRI findings of extramedullary myeloma and to identify its role in the workup of extramedullary myeloma. Materials and Methods This retrospective study was approved by the institutional review board at our tertiary cancer center and was conducted in compliance with HIPAA with waiver of the requirement for informed consent. An electronic search of our radiology database from January 2004 through December 2012 revealed the cases of 156 patients with a diagnosis of MM who underwent crosssectional imaging. Among these 156 patients, 40 patients had extramedullary myeloma either at the time of diagnosis or during the course of disease. MRI was performed for evaluation of extramedullary myeloma or for other indications (with incidental detection of extramedullary myeloma) in 28 patients (15 men, 13 women; mean age, years; range, years) at one or multiple time points. In total, 44 MRI examinations of various sites in the body were studied for these 28 patients. All 28 patients had the diagnosis of MM confirmed by bone marrow biopsy and serum protein electrophoresis. Extramedullary myeloma was confirmed at histopathologic examination for at least one site in all patients. Tirumani et al. TABLE 1: Distribution of Extramedullary Myeloma Lesion Location Contiguous With Bone (n = 44) Noncontiguous With Bone (n = 28) Epidural-paraspinal masses 12 Cervical 2 Thoracic 8 Lumbar 2 Isolated epidural mass 6 Cervical 3 Thoracic 2 Lumbar 1 Isolated paraspinal mass 10 Iliopsoas masses 2 Iliac mass 6 Intracranial, head, neck 9 6 Thorax 1 5 Lung 1 Pleura 2 Mediastinum 1 2 Abdomen 8 Liver 2 Pancreas 3 Adrenal 1 Peritoneal, perirenal 2 Retroperitoneum 5 Subcutaneous 2 MRI Acquisition The 44 MRI examinations included studies of the spine (n = 26); brain, head, and neck (n = 9); abdomen (n = 4); pelvis (n = 4); and chest (n = 1). The MR images were acquired with the following protocols. Images of the cervical, thoracic, and lumbar spine were acquired with axial and sagittal T1-weighted (TR/TE, 516.6/9.6), T2-weighted (TR/TE, 3066/105.7), and STIR sequences and axial and sagittal gadolinium-enhanced T1-weighted (TR/TE, 700/9.4) sequences. Brain, head, and neck images were acquired with axial and coronal T1-weighted spin-echo (TR/TE, 800/13), axial T1-weighted FLAIR (TR/TE, 2800/9.2), axial and coronal T2-weighted FLAIR (TR/TE, 9000/129), axial T2-weighted fast spin-echo (TR/ TE, 4150/76.6), axial and sagittal gadolinium-enhanced T1-weighted, and axial diffusion-weighted sequences. Abdominal images were acquired with axial T2-weighted (TR/TE, 1866/81), T2-weighted fat-suppressed (TR/TE, 2000/98), T1-weighted in- and out-of phase, and unenhanced and gadolinium-enhanced fat-suppressed T1-weighted (TR/ TE, 5.1/2.1) dynamic 3D gradient-recalled echo sequences. Diffusion-weighted images of the abdomen were also available for one patient. Pelvic images were acquired with axial T2-weighted (TR/ TE, 1956/86), axial and coronal STIR, T1-weighted in- and out-of phase, and unenhanced and gadolinium-enhanced fat-suppressed T1-weighted (TR/ TE, 5.8/2.5) sequences. Chest images were acquired with axial T2-weighted (TR/TE, 1869/78), T2-weighted fat-suppressed (TR/TE, 2020/95), inand out-of phase T1-weighted, and unenhanced and gadolinium-enhanced fat-suppressed T1-weighted (TR/TE 5.4/2.3) dynamic 3D gradient-recalled echo sequences. Of the 44 MRI examinations, 32 were performed with gadolinium administration. In patients receiving contrast material, gadolinium was administered at doses of 0.1 mmol/kg body weight up to a maximum dose of 20 ml. Image Analysis Two fellowship-trained radiologists with 7 and 14 years of experience reviewed the images in consensus. For each MR image, the following features of the extramedullary myeloma were noted: location, size (longest diameter for measurable lesions), T1 and T2 signal intensity compared with skeletal muscle (less than, similar to, or greater than skeletal muscle), homogeneity, and degree of enhancement after gadolinium administration (less than, similar to, or greater than skeletal muscle). Also assessed were the presence of adjacent organ, muscle, or neural foramina infiltration; presence of spinal 804 AJR:202, April 2014

3 MRI of Extramedullary Myeloma cord compression; increased T2 signal intensity in the cord; and bone destruction; and concurrent marrow involvement in the other bones. When multiple lesions were present in one organ, it was interpreted as a single site, and the measurement, signal-intensity, and enhancement characteristics of the largest lesion in the organ were used for analysis. Clinical Features and Indications for MRI Of the 44 MRI examinations, six were ordered for further evaluation of findings detected at CT or PET/CT. MRI was required to investigate specific symptoms in the other 38 examinations. The average duration between the diagnosis of MM and MRI was 35 months (range, 0 14 years). Eleven of 28 patients underwent more than one MRI examination, each time for evaluation of a specific symptom. The electronic medical records of all 28 patients were reviewed for date of diagnosis of MM and the type of treatment, including radiotherapy, surgery, and bone marrow transplant. The specific symptom that necessitated MRI and the management after MRI diagnosis of extramedullary myeloma were documented. Results Distribution of Extramedullary Myeloma In the 28 patients, a total of 72 sites of involvement were noted (Table 1). Two radiologic patterns of extramedullary myeloma were recognized in our study: extraskeletal soft-tissue masses present close to the bone (contiguous extramedullary myeloma) (n = 44) and soft-tissue masses distant from bone (noncontiguous extramedullary myeloma) (n = 28). Lesions in both categories were distributed variably throughout the body. The most common sites of involvement of contiguous extramedullary myeloma were paraspinal and epidural (n = 28 [39%]), notably in the thoracic spine (Table 1). In 12 of the 28 sites, there was a combined paraspinal and epidural component. The epidural component was seen to compress the spinal cord in 13 sites; high signal intensity on T2-weighted images of the cord was found in 4 of these 13 sites. Two patterns of paraspinal-epidural involvement were noted (Fig. 1). In the first pattern, seen at 22 sites, the extramedullary myeloma was predominantly centered in the paraspinal location and was associated with adjacent bone destruction and contiguous infiltration of the muscles, retroperitoneum, epidural space, or neural foramina. In the second pattern, seen at six sites, the extramedullary myeloma was predominantly epidural in location and seen as distinct soft-tissue masses (Fig. 2). These epidural masses, though present close to the bone, were Fig. 1 Drawing shows patterns of extramedullary myeloma. Extramedullary myeloma can occur contiguously with bony structures (arrows) or can be noncontiguous with bone (arrowheads). It is presumed that contiguous extramedullary myeloma occurs owing to contiguous extraskeletal extension of myelomatous masses, whereas noncontiguous extramedullary myeloma develops by hematogenous spread of plasma cells. Contiguous extramedullary myeloma occurring around spine can develop predominantly in paraspinal or epidural locations (arrows). associated with minimal bone destruction. The second most common site of contiguous extramedullary myeloma was the intracranial space, and most of these lesions were centered in the anterior cranial fossa (n = 4) (Fig. 3). The iliac and sacral masses (n = 6) were associated with destruction of the adjacent bone in all six cases and infiltration of the sacral plexus in two cases. An anterior mediastinal mass seen in one patient was associated with sternal destruction. The intracranial, head, and neck sites of noncontiguous extramedullary myeloma included the dura (n = 2), masticator space (n = 1), parotid gland (n = 1), anterior mandibular space (n = 1), and cervical (n = 1) and supraclavicular (n = 1) nodes. Dural involvement was noted as diffuse extradural soft tissue with leptomeningeal enhancement in one patient and as dural soft tissue contiguous with a sphenoid mass in another patient. In the abdomen, both the solid viscera (liver, pancreas, adrenal glands) and the retroperitoneum (paraaortic nodes and perirenal nodules) were involved (Fig. 4). The lesions in the liver, pancreas, peritoneal cavity, and retroperitoneum were multiple. In the thorax, the lung (n = 1), pleura (n = 2), and mediastinum (n = 2) were involved. Lung involvement in one patient was seen as a large solitary nodule. Pleural involvement was seen as nodular pleura-based lesions in two patients, one of whom had concurrent pleural effusion. Mediastinal lesions in two patients were seen as a paracardiac node and posterior mediastinalparacardiac masses. Subcutaneous nodules were seen at two sites, one in the anterior abdominal wall and another in the gluteal region. Imaging Features of Extramedullary Myeloma: Comparison of Contiguous and Noncontiguous Extramedullary Myeloma The sites with contiguous extramedullary myeloma were larger and more infiltrative than the sites with noncontiguous extramedullary myeloma (Table 2). The mean size of contiguous extramedullary myeloma was 4.8 cm (range, cm), whereas the mean size of noncontiguous extramedullary myeloma was 3.0 cm (range, cm). This difference was statistically significant (p = 0.001; Student t test). Overall, comparison of the signal intensity of all 72 lesions with skeletal muscle showed that most of the lesions were hypointense to isointense on T1-weighted images (67/72, 93.1%) and isointense to hyperintense on T2-weighted images (62/72, 86.1%). On the T1-weighted images, almost all of the lesions of both contiguous and noncontiguous extramedullary myeloma were hypointense (contiguous, 31/44; noncontiguous, 18/28) or isointense to skeletal muscle (contiguous, 11/44; noncontiguous, 7/28) (Figs. 2 and 4). The lesions hyperintense to muscle on T1-weighted images were predominantly intracranial and head and neck in origin (n = 4) (Fig. 3). On T2- weighted images, all but two contiguous extramedullary myeloma lesions were isointense (n = 7) or hyperintense (n = 35) in relation to muscle. The two lesions were in the paraspinal region and were hypointense. In comparison, 8 of 28 noncontiguous extramedullary myeloma lesions in two patients were hypointense on T2-weighted images (Fig. 4). This difference was statistically sig- AJR:202, April

4 Tirumani et al. A D Fig year-old woman with contiguous extramedullary myeloma in lower thoracic epidural space. A and B, Axial (A) and sagittal (B) T2-weighted MR images of lower thoracic spine show hyperintense posterior epidural mass (arrows) at level of T10 T12 compressing spinal cord. C F, Axial (C and E) and sagittal (D and F) unenhanced (C and D) and gadolinium-enhanced (E and F) T1-weighted MR images show mass (arrows) to be hypointense with homogeneous enhancement. nificant (p = 0.006; Fisher exact test). The distribution of the eight lesions was as follows: peritoneum (n = 1), adrenal gland (n = 1), pancreas (n = 1), mediastinum (n = 2), pleura (n = 1), retroperitoneal node (n = 1), and psoas muscle (n = 1). The other lesions were isointense (n = 4) or hyperintense (n = 16). Almost all of the extramedullary myeloma masses in both categories were homogeneous on T2- weighted images. The exception was four lesions, two in each group, that were heterogeneous masses. None of the lesions exhibited hemorrhage or calcification, although MRI is inferior to CT in detecting calcification. None of the lesions in the abdomen, including the retroperitoneum, had signal-intensity decreases on out-of phase chemical-shift images. On the gadolinium-enhanced images, all sites of extramedullary myeloma had homogeneous enhancement with variable degrees of B E enhancement compared with skeletal muscle (Table 2). Although most of the lesions had mild (contiguous, 15/34; noncontiguous, 4/26) to moderate (contiguous, 16/34; noncontiguous, 14/26) enhancement, intense enhancement was noted at 3 of 34 sites in the contiguous extramedullary myeloma group and 8 of 26 sites in the noncontiguous extramedullary myeloma group. Of note, pancreatic lesions in two patients, liver lesions in one patient, and C F 806 AJR:202, April 2014

5 MRI of Extramedullary Myeloma TABLE 2: MRI Features of Extramedullary Myeloma Feature Contiguous With Bone (n = 44) Noncontiguous With Bone (n = 28) p Size (cm) 4.8 ( ) 3.0 ( ) a T1-weighted images Hypointense Isointense 11 7 Hyperintense 2 3 T2-weighted images Hypointense b Isointense 7 4 Hyperintense Gadolinium-enhanced images Mild 15 4 Moderate Intense 3 8 a Student t test. Fisher exact test. retroperitoneal lesions in two patients were hypervascular in the arterial phase and had persistent enhancement in the venous and delayed phases (Fig. 4). Diffusion-weighted images were available for a total of nine sites in the two groups combined. All the lesions had marked diffusion restriction, the apparent diffusion coefficient values ranging between 0.25 and 0.98 (mean, 0.50). Clinical Features and Outcome of MRI Of the 28 patients, nine had extramedullary myeloma at presentation. In the others, extramedullary myeloma developed during the subsequent disease course. In seven patients extramedullary myeloma developed as a relapse after bone marrow transplant at variable time intervals (6 months 11 years). MRI was required to investigate specific symptoms in 38 examinations. The most common symptom necessitating MRI was back, buttock, neck, or hip pain (n = 18) at presentation or during the course of disease. CNS symptoms such as leg weakness, facial numbness, headache, proptosis, diplopia, and altered mental status prompted 14 examinations. Of these 14, three spinal MRI examinations were ordered to evaluate spinal cord compression by epidural masses due to leg weakness. The other indications included palpable forehead mass, progressive generalized weakness, epigastric pain, elevated liver function test results, chest pain, hoarseness, and palpable masses. After MRI, 14 of 28 patients were treated with local radiotherapy for prompt relief of the mass effect of the extramedullary myeloma in the spine and head. Surgery was performed before radiotherapy in 5 of 28 patients. Discussion MRI has a multidimensional role in the management of MM because it helps in initial diagnosis, staging, clarifying ambiguous skeletal survey or PET/CT findings, disease monitoring, assessing treatment response, and workup of complications [8]. In the evaluation of complications of MM, MRI is useful as a problem-solving tool. National Comprehensive Cancer Network guidelines [7] call for MRI in cases of suspected vertebral compression fractures. The most common indications for ordering MRI in our study population (n = 32) were neurologic symptoms such as back pain, leg weakness, facial numbness, proptosis, diplopia, and altered mental status. The MRI features of MM are widely described and include normal marrow pattern, micronodular salt-and-pepper pattern, and focal and diffuse patterns [1, 9]. However, literature on the MRI features of extramedullary myeloma is scant [10, 11]. To our knowledge, our study includes the largest series of cases in which the MRI features of extramedullary myeloma A Fig year-old man with contiguous extramedullary myeloma in right sphenoid region. A, Coronal fat-suppressed T2-weighted MR image of skull base shows hyperintense mass (arrows) in sphenoid region. B and C, Coronal T1-weighted unenhanced non fat-suppressed (B) and gadolinium-enhanced fat-suppressed (C) MR images of skull base show mass (arrows) to be homogeneously hyperintense with homogeneous enhancement. B C AJR:202, April

6 Tirumani et al. are described. We also report for the first time two radiologic patterns of extramedullary myeloma: contiguous and noncontiguous extramedullary myeloma, which though known in oncology have not been described in the radiology literature. In a study of 12 patients, Ooi et al. [11] found that extramedullary myeloma plasmacytomas tend to be isointense on T1-weighted images and isointense to hyperintense relative to muscle and white matter on T2-weighted images and exhibit mild to marked enhancement. We made similar observations, but there were some differences. We observed two patterns of extramedullary myeloma: contiguous with bone and not contiguous with bone. The contiguous lesions were larger than the noncontiguous lesions. The contiguous and noncontiguous extramedullary myelomas did not differ significantly in T1 signal-intensity characteristics in that they were predominantly homogeneously isointense to hypointense relative to skeletal muscle (Fig. 2). Though a substantial number of lesions in both groups were hyperintense relative to skeletal muscle on T2- weighted images (62/72, 86.1%), a number of noncontiguous lesions (8/28, 28.6%) were hypointense, and this difference was statistically significant (Fig. 4). This T2 hypointensity, which is a hallmark of lymphoma, may be explained in part by the biologic similarity of extramedullary myeloma, especially noncontiguous extramedullary myeloma, and lymphoma. The T2 hypointensity of extramedullary myeloma has been described [12]. The presence of T2 hypointense lesions in only 2 of 28 patients in our study and the concurrence of T2 hyperintense lesions at other sites in these two patients make it difficult to draw definite conclusions. The 28 noncontiguous extramedullary myeloma masses in our study are presumed to have arisen from hematogenous dissemination, abdominal viscera being the most common site in our study. Contiguous extramedullary myeloma (n = 44/72, 61%) was the dominant variety in our study, and most of these masses presented in paraspinal and epidural locations. To our knowledge, there have been fewer than 20 individual case reports of spinal epidural extramedullary myeloma since 1950 [13, 14]. Predominantly epidural masses constituted six cases in our study (Fig. 2). All of the contiguous extramedullary myeloma lesions had their epicenter away from the bone, even when associated with bone destruction. The exact pathogenesis of contiguous extramedullary myeloma is unclear, but we presume that it arises from contiguous extraskeletal extension of myelomatous masses. The origin of isolated epidural extramedullary myeloma is also speculative. Few hypotheses have been proposed and include paraspinal nodes extending along the neural foramen and transformation of epidural lymphoid tissue. The development of paraspinal and epidural masses in extramedullary myeloma is clinically important because of the risk of pathologic vertebral fractures and cord compression. MRI has a specific role in this setting because it facilitates evaluation of the spinal cord and can help determine the extent of the epidural disease to aid in treatment planning. Cord compression was noted in 13 of 28 lesions in our study, though high T2 signal intensity suggestive of cord edema was noted in only 4 of 13 lesions. From the management point of view, the presence of epidural masses often A C necessitates prompt institution of steroid treatment, radiotherapy, or surgery. Extramedullary myeloma has a predilection for the nasopharynx and sinuses [15]. Relapse in the intracranial region is uncommon, constituting fewer than 1% of relapses after autologous stem cell transplant [16]. A striking feature of the intracranial extramedullary myeloma in our study was the homogeneous isointensity to hyperintensity of the masses relative to skeletal muscle on T1-weighted images (Fig. 3). None of the lesions was hypointense on T2-weighted images to suggest hemorrhage. The cause of this finding is unknown but could be related to the high cellularity of extramedullary myeloma. The conventional differential diagnoses for malignant brain tumors, which are isointense to hyperintense on T1-weighted images, include meningioma, metastasis from melanoma, and prima- Fig year-old man with noncontiguous extramedullary myeloma in left perirenal space. A and B, Axial T2-weighted (A) and fat-suppressed unenhanced T1-weighted (B) MR images show T1 and T2 hypointense left posterior perirenal nodule (arrow). C and D, Axial fat-suppressed dynamic gadolinium-enhanced T1-weighted MR images in arterial (C) and venous (D) phases show moderate arterial and persistent progressive venous enhancement of nodule (arrow). B D 808 AJR:202, April 2014

7 MRI of Extramedullary Myeloma ry diffuse meningeal melanomatosis [17]. We found that intracranial and head extramedullary myeloma exhibited marked diffusion restriction on diffusion-weighted images, which is also indicative of the high cellularity of extramedullary myeloma. The indications for abdominal MRI in our study were evaluation of epigastric pain and assessment of FDG-avid lesions at PET/CT. Noncontiguous extramedullary myeloma in the abdomen was notable for its multiplicity in our study. As in other sites, all of the lesions in the abdomen were homogeneous. Hepatic and pancreatic involvement has been found in 30% and 4% of cases of MM at autopsy and can be focal (unifocal or multifocal) or diffuse [12, 18]. The hepatic lesions in two of our patients were multifocal, were hypointense in relation to skeletal muscle on T1-weighted images, and were mildly hyperintense on T2-weighted images. On gadolinium-enhanced images, the lesions in one patient were hypervascular, and in the other patient they were hypovascular, nevertheless exhibiting gradual enhancement in the venous and delayed phases. Overall the imaging findings were nonspecific, a finding reported earlier by Ooi et al. [11]. The pancreatic lesions were notable for hypervascularity in two patients. Hypervascularity of visceral extramedullary myeloma has been reported and has been attributed to the hypervascular nature of MM in general due to increased angiogenesis [12, 19]. The pancreatic lesions are indeed difficult to differentiate from hypervascular metastases from renal cell carcinoma, melanoma, and neuroendocrine tumors [12, 20]. Peritoneal nodules were multiple and isointense to hypointense on T2-weighted images, mimicking lymphoma, as reported in earlier literature [20, 21]. Renal extramedullary myeloma occurs as large solid renal masses or perirenal nodules [22, 23]. Perirenal nodules were seen in two patients in our study. Multiple homogeneous retroperitoneal nodes of variable signal intensity and moderate to marked enhancement was the most common pattern of retroperitoneal involvement, indistinguishable from that of lymphoma [20, 21] (Fig. 4). The lesions in the thorax were isointense to hypointense in relation to skeletal muscle on T1-weighted images. Both the mediastinal lesions and one pleura-based nodule were mildly hypointense on T2-weighted images. Pleural effusion was seen in one patient with a pleural nodule. The lung nodule was indistinguishable from a metastatic nodule. Subcutaneous nodules did not differ from those at other sites of disease in being hypointense on T1-weighted images and hyperintense on T2- weighted images. In a patient with previously undiagnosed MM, extramedullary myeloma is unusual (< 20%). In patients with MM, the differential diagnoses for soft-tissue masses include extramedullary myeloma, amyloidosis, and second malignancy, either primary or metastatic. After a transplant, T2 hypointense or T1 hyperintense masses exhibiting moderate to marked enhancement and highly restricted diffusion can suggest the possibility of extramedullary myeloma. In our experience, the role of MRI in extramedullary myeloma is mainly to evaluate specific symptoms and ambiguous findings seen at CT or PET/CT. MRI was also helpful in treatment planning for patients undergoing radiotherapy or surgery (n = 19/28). Our study had limitations, including the retrospective design and small sample size. The conclusions drawn from this study may therefore have to be evaluated in studies with large sample sizes and cytogenetic correlation. Because the study was confined to MRI, the outcome and follow-up of extramedullary myeloma were not evaluated. Histopathologic evaluation of most of the lesions, especially paraspinal and epidural masses, was not performed because the diagnosis of extramedullary myeloma was based on the clinical findings, confirmation of MM with blood tests and bone marrow biopsy, and exclusion of secondary malignancies. We did not evaluate the correlation between the type of myeloma and the MRI appearance because of the small sample size. We also did not evaluate the role of whole-body MRI in screening and staging of MM. Conclusion Extramedullary myeloma is an unusual entity that has nonspecific imaging features at MRI. Nevertheless, knowledge of its imaging features may help radiologists to suspect it in the appropriate clinical scenario. We describe two radiologic patterns of extramedullary myeloma: contiguous with bone and noncontiguous with bone. Extramedullary myeloma is usually homogeneous and hypointense compared with skeletal muscle on T1-weighted images and hyperintense on T2-weighted images and exhibits mild to intense enhancement. Contiguous extramedullary myeloma most commonly occurs as a large mass in paraspinal and epidural locations. Some noncontiguous extramedullary myelomas mimic lymphoma because of their homogeneous and low T2 signal intensity. MRI is a problemsolving tool in extramedullary myeloma and helps in treatment planning. References 1. Hanrahan CJ, Christensen CR, Crim JR. Current concepts in the evaluation of multiple myeloma with MR imaging and FDG PET/CT. Radio- Graphics 2010; 30: Kumar SK, Rajkumar SV, Dispenzieri A, et al. Improved survival in multiple myeloma and the impact of novel therapies. Blood 2008; 111: Usmani SZ, Heuck C, Mitchell A, et al. Extramedullary disease portends poor prognosis in multiple myeloma and is over-represented in high-risk disease even in the era of novel agents. Haematologica 2012; 97: Varettoni M, Corso A, Pica G, Mangiacavalli S, Pascutto C, Lazzarino M. Incidence, presenting features and outcome of extramedullary disease in multiple myeloma: a longitudinal study on 1003 consecutive patients. Ann Oncol 2010; 21: Bladé J, Fernández de Larrea C, Rosiñol L, Cibeira MT, Jiménez R, Powles R. Soft-tissue plasmacytomas in multiple myeloma: incidence, mechanisms of extramedullary spread, and treatment approach. J Clin Oncol 2011; 29: Bladé J, de Larrea CF, Rosiñol L. Extramedullary involvement in multiple myeloma. Haematologica 2012; 97: National Comprehensive Cancer Network website. NCCN guidelines: Multiple myeloma. org/professionals/physician_gls/f_guidelines.asp# site. Accessed November 24, Shortt CP, Gleeson TG, Breen KA, et al. Wholebody MRI versus PET in assessment of multiple myeloma disease activity. AJR 2009; 192: Stäbler A, Baur A, Bartl R, Munker R, Lamerz R, Reiser MF. Contrast enhancement and quantitative signal analysis in MR imaging of multiple myeloma: assessment of focal and diffuse growth patterns in marrow correlated with biopsies and survival rates. AJR 1996; 167: Birjawi GA, Jalbout R, Musallam KM, Tawil AN, Taher AT, Khoury NJ. Abdominal manifestations of multiple myeloma: a retrospective radiologic overview. Clin Lymphoma Myeloma 2008; 8: Ooi GC, Chim JC, Au WY, Khong PL. Radiologic manifestations of primary solitary extramedullary and multiple solitary plasmacytomas. AJR 2006; 186: Kazama T, Ng CS, Giralt SA. Multiphasic CT and MRI appearances of extramedullary multiple myeloma involving the stomach, pancreas, and bladder. Clin Imaging 2005; 29: Lourbopoulos A, Ioannidis P, Balogiannis I, Stavrinou P, Koletsa T, Karacostas D. Cervical epidural plasmacytoma presenting as ascending AJR:202, April

8 Tirumani et al. paraparesis. Spine J 2011; 11:e1 e4 14. Avadhani A, Shetty AP, Rajasekaran S. Isolated extraosseous epidural myeloma presenting with thoracic compressive myelopathy. Spine J 2010; 10:e7 e Leake PA, Coard KC, Plummer JM. Extramedullary plasmacytoma of the pancreas as an uncommon cause of obstructive jaundice: a case report. J Med Case Rep 2009; 3: Waterhouse DF, Moloney GA, Gargoum FS, Hayden PS, O Gorman T. Extramedullary relapse of IgA-lambda myeloma after recent bortezomib therapy: a case report. Cases J 2009; 2:7456 FOR YOUR INFORMATION 17. Ginat DT, Meyers SP. Intracranial lesions with high signal intensity on T1-weighted MR images: differential diagnosis. RadioGraphics 2012; 32: Oshima K, Kanda Y, Nannya Y, et al. Clinical and pathologic findings in 52 consecutively autopsied cases with multiple myeloma. Am J Hematol 2001; 67: Patlas M, Khalili K, Dill-Macky MJ, Wilson SR. Spectrum of imaging findings in abdominal extraosseous myeloma. AJR 2004; 183: Philips S, Menias C, Vikram R, Sunnapwar A, Prasad SR. Abdominal manifestations of extraosseous myeloma: cross-sectional imaging spectrum. J Comput Assist Tomogr 2012; 36: Hall MN, Jagannathan JP, Ramaiya NH, Shinagare AB, Van den Abbeele AD. Imaging of extraosseous myeloma: CT, PET/CT, and MRI features. AJR 2010; 195: Monill J, Pernas J, Montserrat E, et al. CT features of abdominal plasma cell neoplasms. Eur Radiol 2005; 15: Moulopoulos LA, Granfield CA, Dimopoulos MA, Kim EE, Alexanian R, Libshitz HI. Extraosseous multiple myeloma: imaging features. AJR 1993; 161: This article is available for CME and Self-Assessment (SA-CME) credit that satisfies Part II requirements for maintenance of certification (MOC). To access the examination for this article, follow the prompts associated with the online version of the article. 810 AJR:202, April 2014