Evaluation of the diagnostic utility of spinal magnetic resonance imaging in axial spondylarthritis.
код для вставкиСкачатьARTHRITIS & RHEUMATISM Vol. 60, No. 5, May 2009, pp 1331–1341 DOI 10.1002/art.24493 © 2009, American College of Rheumatology Evaluation of the Diagnostic Utility of Spinal Magnetic Resonance Imaging in Axial Spondylarthritis A. N. Bennett,1 A. Rehman,2 E. M. A. Hensor,3 H. Marzo-Ortega,3 P. Emery,3 and D. McGonagle4 were of low diagnostic utility for SpA, but >3 RLs (likelihood ratio [LR] 12.4) and severe RLs (LR infinite) in younger subjects were highly diagnostic of SpA. Posterior element BME lesions of mild or moderate grade were also highly diagnostic of SpA (LR 14.5). The most common diagnostic confusion was between SpA and DA, since both had RLs present and the presence/ absence of degenerative discs did not change the diagnostic assessment. Conclusion. This study confirms the high diagnostic utility of MRI in axial SpA, with severe or multiple RLs evident on MRI being characteristic in younger patients and mild/moderate posterior element lesions being specific for SpA. However, MRI lesions previously considered to be characteristic of SpA could also be found frequently in patients with DA and patients with malignancy, and therefore such lesions should be interpreted with caution, particularly in older patients. Objective. Magnetic resonance imaging (MRI) is increasingly used for the diagnosis of axial spondylarthritis (SpA), but it is unknown whether characteristic lesions are actually specific for SpA. This study was undertaken to compare MRI patterns of disease in active SpA, degenerative arthritis (DA), and malignancy. Methods. Fat-suppressed MRI of the axial skeleton was performed on 174 patients with back pain and 11 control subjects. Lesions detected by MRI, including Romanus lesions (RLs) and end-plate, diffuse vertebral body, posterior element, and spinous process bone marrow edema (BME) lesions, were scored in a blinded manner. An imaging diagnosis was given based on MRI findings alone, and this was compared with the goldstandard treating physician’s diagnosis. Results.The physician diagnosis was SpA in 64 subjects, DA in 45 subjects, malignancy in 45 subjects, other diagnoses in 20 subjects, and normal in 11 subjects. There was 72% agreement between the imaging diagnosis and physician diagnosis. End-plate edema, degenerative discs, and RLs were frequently observed in patients with any of the 3 major diagnoses. Single RLs In recent years, magnetic resonance imaging (MRI) has become established as a diagnostic tool for spondylarthritis (SpA), particularly in early sacroiliac (SI) joint disease. This is principally because the diagnostic radiographic changes in the SI joints and spine usually develop slowly (1). Fat-suppressed MRI enables identification of osteitis at disease presentation and has the potential to transform the diagnosis of axial SpA, since much of the axial pathology is related to a diffuse perifibrocartilage osteitis, either in the SI joints or at the entheses in the spine (2–4). Many features of SpA that have been identified on MRI, including Romanus lesions (RLs), end-plate lesions, diffuse vertebral body lesions, posterior element (facet and pedicle) lesions, and spinous process bone marrow edema (BME) lesions, are recognized as characteristic of SpA (5–7). It is generally assumed that these MRI patterns of disease can be used to distinguish SpA Dr. Hensor’s work is supported in part by an ARC program grant. 1 A. N. Bennett, MBChB, MRCP, MSc, MFSEM: Leeds Institute of Molecular Medicine and Chapel Allerton Hospital, Leeds, and Defence Medical Rehabilitation Centre–Headley Court, Epsom, UK; 2A. Rehman, MBBS, MRCP, FRCR: American Hospital Dubai, Oud Metha, Dubai, United Arab Emirates; 3E. M. A. Hensor, PhD, H. Marzo-Ortega, MRCP, PhD, P. Emery, MA, MD, FRCP: Leeds Institute of Molecular Medicine and Chapel Allerton Hospital, Leeds, UK; 4D. McGonagle, PhD, FRCPI: Leeds Institute of Molecular Medicine and Chapel Allerton Hospital, Leeds, UK, and Merlin Park Hospital, Galway, Ireland. Address correspondence and reprint requests to D. McGonagle, PhD, FRCPI, Academic Section of Musculoskeletal Disease, Leeds Institute of Molecular Medicine, Chapel Allerton Hospital, Chapeltown Road, Leeds LS7 4SA, UK. E-mail: d.g.mcgonagle@ leeds.ac.uk. Submitted for publication August 21, 2008; accepted in revised form February 9, 2009. 1331 1332 BENNETT ET AL from other diseases. Indeed, these lesions are being used to diagnose axial SpA early in its onset (3,8) and might influence early initiation of biologic therapy; for this purpose, several systems of scoring spinal BME lesions have been established (6,9–12). However, the specificity of these spinal BME lesions for SpA and their diagnostic utility for SpA have not been confirmed. Despite this, axial imaging continues to assume increasing importance in SpA. Furthermore, acute inflammation in axial SpA affects only the spine and spares the SI joints in 23% of cases (13), so there is a need to better define the role of MRI in this group. However, distinguishing spinal MRI patterns of different diseases may not be simple. We have previously studied distal interphalangeal (DIP) joint disease in 2 distinctly different conditions, psoriatic arthritis (PsA) and osteoarthritis (OA), and found that MRI patterns of disease were difficult to differentiate, because the entheseal and bone-related changes were similar (14). We therefore assessed the pattern of MRI lesions typically noted in SpA across 3 major causes of chronic back pain, the clinically confirmed European Spondylarthropathy Study Group (ESSG) criteria for SpA (15), degenerative spinal arthritis (DA), and spinal malignancy/metastases, in comparison with a normal control group of subjects without back pain. A wide age group was studied to capture all ages at presentation of SpA (16–18), and particularly to capture the 30–50-year-old age group, in whom limited data suggest that early degenerative changes on MRI can be mistaken for axial SpA (19). The diagnostic utility of these characteristic changes in SpA was analyzed. In particular, MRI patterns of disease were compared between conditions to ascertain whether distinct anatomic patterns of disease exist, and whether, as highlighted in the research agenda of the Outcome Measures in Rheumatology Clinical Trials (OMERACT) 7 (20), BME is specific for the inflammation in SpA. PATIENTS AND METHODS Cohort. Patients were recruited from Calderdale Royal Hospital (CRH), West Yorkshire, UK. Ethics committee approval was obtained, and all living patients gave their written informed consent. Potential patients were identified by clinical radiology staff, via the Patient Archiving and Communication Systems at CRH, as having had MRI of the whole spine with or without imaging of the SI joints for back pain in the last 5 years. Patients who met these criteria were invited to participate in the study. A control group of subjects with no back pain was also recruited. The diagnosis that was given by the treating physician was considered the gold standard. The gold-standard physician diagnosis was based on a history review, physical examination, nonradiologic and radiologic investigations, histologic findings (when available), and clinical outcome data. Patients’ case records, with the benefit of up to 5 years of outcome data, were then evaluated to establish whether the treating physician’s diagnosis was correct. In the SpA cohort (n ⫽ 64), all patients met the ESSG criteria for SpA (15) prior to MRI scanning. Specific diagnoses in this group were ankylosing spondylitis (AS) in 20 patients, undifferentiated SpA in 17 patients, inflammatory bowel disease– related SpA in 14 patients, PsA in 9 patients, and reactive arthritis in 4 patients, with a mean disease duration of 8.5 years (AS patients, mean disease duration 15.2 years). Ninety-four percent of the patients with AS and 68% of all SpA patients were HLA–B27 positive. All SpA patients had clinically active disease, and MRI scans had been performed to assess ongoing disease activity. Three patients were receiving anti–tumor necrosis factor therapy at the time of the scan but still had clinically active disease. Sample size. Power calculations were performed using specificity and sensitivity nomograms (21). The prevalence of SpA in the present study cohort was 35%. A sample size of 150 allowed an accurate evaluation of MRI lesions in the diagnosis of SpA, with measurement of specificity and sensitivity at the 95% level with 6% precision at 95% confidence levels. MRI scoring. A total of 185 patients had MR images of the whole spine, comprising sagittal T1 and STIR (fatsuppressed) sequences. In addition, MR images of the SI joints, comprising coronal/oblique T1 and STIR sequences, were obtained in 89 patients (48%), of whom 91% had SpA and 26% had other diagnoses. This reflects standard clinical practice, since many patients with DA or malignant spinal disease do not have the SI joints scanned. Whole-spine MRI scan protocols were standardized across the different patient groups. All MR scans were scored by consensus by 2 experienced observers (AR and DM), who were blinded to the patients’ details and clinical characteristics. The Leeds Scoring System (6,12) was used to score spinal BME as well as BME of the SI joints on STIR sequences. BME was defined, according to the OMERACT definitions (22,23), as low signal intensity on the T1 sequence and high/ intermediate bone marrow signals with irregular contour on the STIR image, when compared with normal bone marrow signals coming from the sacral interforaminal area in the SI joint scans or the center of an unaffected vertebra in the spinal scans, similar to the Spondyloarthritis Research Consortium of Canada definition (9). In the SI joints, a semiquantitative global scale was used to score the severity of BME according to the extent of the quadrant affected, as follows: grade 0 ⫽ normal, grade 1 ⫽ ⬍25%, grade 2 ⫽ 25–75%, or grade 3 ⫽ ⬎75%. This system has been used in previous studies (6,12) and, as reported in OMERACT 7 (20,24), is considered equivalent to alternative scoring systems. In addition to the scoring of active BME lesions, structural changes, as shown by the presence or absence of erosions and ankylosis, were also recorded, with erosions identified as subchondral defects with a low signal on T1-weighted images, and ankylosis defined as complete bridging of the SI joints with total loss of joint space on T1 sequences. SPINAL MRI DIAGNOSIS OF AXIAL SpA For the spine, the Leeds Scoring System was used to identify different patterns of vertebral lesions, since some patterns of acute vertebral inflammation are thought to also occur frequently in DA (9,19). The scoring system was modified to grade the severity of the lesions, since it was considered important to use a semiquantitative system rather than a dichotomous system to grade the severity of spinal lesions, because severity of BME lesions in the SI joints has been shown to play a role in the early diagnosis of AS (3). We decided to modify the Leeds Scoring System rather than use other available scoring systems (9–11), since other systems do not include posterior elements (25), which is an area that seems to be highly relevant to SpA, as has also been noted by other authors (26,27). A training session was arranged prior to these assessments, in which multiple scans were reviewed to recognize different examples of the wide range of spinal pathologic findings described in the literature for SpA, malignancy, and DA (5–7,28–33). Readers discussed the scans and agreed on definitions. All lesions were identified as those displaying a high signal on STIR sequences, as described above. Specific pathologic findings were identified for each type of lesion. RLs were defined as BME lesions arising from any corner of any vertebrae from the inferior C2 to superior S1, extending toward the center of the vertebrae but not occupying more than 50% of the end plate. End-plate lesions were those arising from either the superior or inferior end plates from the inferior C2 to superior S1, extending across more than 50% of the end plate beyond the center of the vertebrae. Posterior element lesions were those arising from either the pedicle or facet joint from C2 to L5, with extension toward either the vertebral body or the spinous process. Spinous process lesions were those arising from the spinous process of any vertebrae from C2 to L5, with possible extension toward the facet joint/pedicle or toward the intraspinous ligaments. Diffuse vertebral body edema lesions were identified as those showing a relatively homogeneous high signal within the whole of the vertebral body from any vertebrae from C2 to S1. All spinal lesions were graded using a semiquantitative scale, ranging from 0 to 3 (0 ⫽ normal, 1 ⫽ mild, 2 ⫽ moderate, 3 ⫽ severe) according to the intensity and extent of the BME signal affecting the anatomic structure involved, i.e., quadrants of vertebrae for RLs, half vertebrae for end-plate lesions, whole vertebrae for diffuse vertebral edema lesions, facet and pedicle for posterior element lesions, and spinous process for spinous process lesions. This grading system (range 0–3 for mild–severe) is very similar to that used in the Berlin Scoring System (34) and the ASpireMRI BME scoring system (10), the difference being that the modifications in the Leeds Scoring System allowed additional scoring of posterior elements, including spinous processes, and allowed differentiation of vertebral lesions, including RLs, end-plate lesions, or diffuse vertebral edema lesions. Other frequently reported lesions were those typical of malignancy, i.e., diffuse patchy uptake with low signal on T1 and medium/high signal on STIR, bone destruction, expansile/lytic/sclerotic lesions, or soft-tissue masses (30). Typical degenerative changes in the degenerative discs or end-plate lesions were also recorded (31,35). In those cases where both the spine images and SI joint images were available, they were scored separately, reflecting the fact that in the investigation of back pain across different 1333 specialties in clinical practice, it is common for spinal MRI to be performed without SI joint MRI sequences. Diagnoses based on spinal images were given separately from those based on SI joint scans, followed by an overall diagnosis that was determined on the basis of the 2 combined, when SI joint scans were also available. The overall imaging diagnosis was compared with the gold-standard treating physician’s diagnosis, and the frequency and pattern of lesions on MRI were compared with the gold-standard physician diagnosis to assess diagnostic utility. Reliability of scoring. The reliability of the MRI scoring was assessed for both the SI joints and the spine separately, on 25 scans that were read at 2 time points, 1 week apart. Agreement on overall diagnosis was 100% for readings of the SI joint scans and 100% for readings of the spine scans (overall quadratic-weighted kappa [Kw] ⫽ 1.00 for both). The reliability of scoring specific SI joint lesions and agreement between the grade of the SI joint lesion was almost perfect (36), with an overall exact agreement of 93% (Kw ⫽ 0.95). Identification of chronic SI joint lesions on T1 scans had excellent agreement, with 100% agreement (K ⫽ 1.00) for both erosion and ankylosis. In the spine, agreement between different grades of different types of lesions (36) was also substantial (prevalence-adjusted bias-adjusted kappa, 0.68 for RLs, 0.90 for end-plate lesions, 0.89 for posterior element lesions, 0.83 for spinous process lesions, and 0.90 for diffuse edema). Statistical analysis. Cross-table analysis was used to calculate the sensitivity, specificity, and likelihood ratios (LRs) of specific lesions in relation to specific diagnoses, using the treating physician’s diagnosis as the gold standard. LRs are regarded as one of the best ways to measure and express diagnostic accuracy (37). All statistical analyses were carried out in SPSS, version 15.0.1.1 (SPSS, Chicago, IL). RESULTS Characteristics of the subjects. One hundred eighty-five participants were recruited. The goldstandard physician diagnosis was SpA in 64 patients (35%), DA in 45 (24%), and metastases in 45 (24%), while 11 patients (6%) had normal clinical findings and 20 (11%) were diagnosed as having other conditions, including tuberculosis abscess, multiple hemangiomas, osteoporotic fractures, and radiation osteonecrosis. The mean ⫾ SD age of the whole cohort was 52.5 ⫾ 17.3 years, ranging from 19 years to 91 years, of whom 89 (48%) were under age 50 years (47 patients with SpA, 18 with DA, 9 with malignancy, 4 with other diagnoses, and 11 with normal findings) and 98 (53%) were in the age range of 30–60 years. As expected, SpA patients were significantly younger (mean ⫾ SD 41.3 ⫾ 12.7 years) than those with a gold-standard diagnosis of DA (61.9 ⫾ 14.3 years; P ⬍ 0.05) or spinal malignancy (64.4 ⫾ 13.1 years; P ⬍ 0.05). Imaging diagnosis versus gold-standard diagnosis. When all lesions were collated and a global impression assigned, there was 72% agreement between 1334 Figure 1. Characteristic spondylarthritis (SpA) Romanus lesions (RLs) in different diagnoses. A, Grade 3 RLs in the inferior-anterior corner of T9 and grade 2 RLs in the anterior-superior of T10 (arrow for both) in a patient with SpA. B, Grade 2 RLs in the inferior-anterior of L3 (arrow) and a small annular tear in the L3–L4 disc in a patient with degenerative spinal disease. C, Grade 2 RLs in the posteriorinferior corner of L3 (arrow) in a patient with metastatic breast cancer. the imaging diagnosis and gold-standard diagnosis. The agreement for a diagnosis of SpA alone was 75%, while for DA it was 62%. The agreement for malignancy was highest, at 91%. The most common discrepancy in diagnosis was between SpA and DA, since both conditions commonly had RLs, end-plate lesions, and degenerative discs present. Degenerative discs were a common finding across the whole cohort (38%), and therefore this feature was of little diagnostic utility. Incorrect diagnosis based on MRI findings alone. In the SpA group, 16 subjects (25%) with a gold-standard physician diagnosis of SpA in the study (all fulfilling the ESSG criteria) were given another diagnosis on the basis of the MRI scan alone. Twelve of the 16 subjects were misdiagnosed as having DA (mean age 41 years), based on MRI evidence of RLs (6 of 12), end-plate lesions (9 of 12), and degenerative discs (9 of 12). These lesions were mostly mild and frequently occurred in isolation, often adjacent to degenerative discs in the lower cervical and lumbar spine. Three subjects had normal findings on the MRI scans despite active disease, which is consistent with cases described in previous reports (10,13), and 1 patient was misdiagnosed as having malignancy based on the MRI scans. In the DA group, 17 patients were given a diagnosis of SpA based on MRI alone. The median age BENNETT ET AL of this group was 59 years (range 26–79 years), and 88%, 53%, and 29% of the patients had RLs, end-plate lesions, and posterior element lesions present, respectively. The trend, both in SpA patients and in DA patients, was for an increase in the number of BME lesions with increasing age. The median lesion count per SpA patient was 4.86 in those under age 50 years (n ⫽ 44) and 5.5 in those over age 50 years (n ⫽ 19). In the DA patients, the median lesion count was 3.3 in patients under age 50 years (n ⫽ 12) and 5.2 in those over age 50 years (n ⫽ 33). In the spinal malignancy group, RLs were frequently present on MRI; however, these were rarely found in isolation and almost all cases also had other abnormalities, such as expansile, lytic, or sclerotic lesions in a single or multiple vertebrae or adjacent soft tissues, clearly suggesting a diagnosis of malignancy. Nevertheless, 3 patients with confirmed metastatic disease (all secondary to breast carcinoma) were given an imaging diagnosis of SpA because of the limited distribution of early metastatic disease at a stage when it was still nondestructive (Figures 1C and 2A). Figure 2. Incorrect magnetic resonance imaging (MRI) diagnoses made on the basis of characteristic spondylarthritis (SpA) lesions evident on MRI. A, Imaging diagnosis of SpA on the basis of Romanus lesions (RLs) (grade 2) in the anterior-inferior border of T4 and anterior-superior border of T5 (arrow for both), spinous process bone marrow edema lesions at T3–T5 (grades 3, 2, and 2, respectively) (triple arrow), and diffuse vertebral body edema at T9 (grade 2) (arrowhead). The actual diagnosis is spinal bone metastases. B, Imaging diagnosis of SpA on the basis of classic MRI evidence of coalescent RLs anterior-superior and anterior-inferior to L4 (grade 1) (arrow) on 3 consecutive slices, adjacent to well-hydrated healthy intervertebral discs. The actual diagnosis is degenerative arthritis (DA). C, Imaging diagnosis of DA on the basis of a large RL (grade 3) anterior-inferior to L5 (arrow) in a classic “degenerative” location at L5–S1, adjacent to a degenerative disc with a small anterior annular tear. The actual diagnosis is psoriatic arthritis/SpA. SPINAL MRI DIAGNOSIS OF AXIAL SpA 1335 Table 1. Sensitivity, specificity, and positive likelihood ratio (LR) for different grades and amounts of Romanus lesions (RLs) and different grades of posterior element lesions in the diagnosis of spondylarthritis (SpA), degenerative spinal arthritis (DA), and spinal malignancy* MRI characteristic, diagnosis Sensitivity (95% CI) Specificity (95% CI) 0.67 (0.53–0.79) 0.63 (0.47–0.77) 0.39 (0.25–0.54) 0.56 (0.47–0.65) 0.52 (0.43–0.61) 0.44 (0.35–0.53) 1.5 (1.2–2.0) 1.3 (1.0–1.8) 0.7 (0.5–1.0) 0.45 (0.32–0.59) 0.37 (0.23–0.53) 0.11 (0.04–0.25) 0.81 (0.73–0.88) 0.76 (0.67–0.83) 0.67 (0.58–0.75) 2.5 (1.5–4.0)† 1.5 (0.9–2.5) 0.4 (0.2–0.8) 0.33 (0.21–0.47)† 0.05 (0.01–0.18) 0.00 (0.00–0.10) 0.97 (0.92–0.99)† 0.85 (0.78–0.91) 0.83 (0.75–0.89) 12.4 (3.8–40.4)† 0.3 (0.1–1.4) 0.00 0.16 (0.08–0.29) 0.12 (0.05–0.27) 0.18 (0.09–0.33) 0.89 (0.81–0.94) 0.87 (0.79–0.92) 0.89 (0.82–0.94) 1.4 (0.7–3.2) 0.9 (0.4–2.3) 1.6 (0.7–3.6) 0.11 (0.05–0.23) 0.00 (0.00–0.11) 0.00 (0.00–0.10) 1.00 (0.96–1.00) 0.95 (0.90–0.98) 0.95 (0.89–0.98) 0.29 (0.18–0.43) 0.12 (0.05–0.27) 0.20 (0.10–0.36) 0.87 (0.79–0.92) 0.80 (0.71–0.86) 0.82 (0.74–0.88) 2.2 (1.2–4.1) 0.6 (0.3–1.5) 1.2 (0.6–2.3) 0.16 (0.08–0.29) 0.10 (0.03–0.24) 0.20 (0.10–0.36) 0.88 (0.80–0.93) 0.85 (0.78–0.91) 0.89 (0.82–0.94) 1.3 (0.6–2.9) 0.7 (0.2–1.8) 1.8 (0.9–3.9) 0.13 (0.06–0.25) 0.02 (0.00–0.14) 0.00 (0.00–0.10) 0.99 (0.94–1.00) 0.95 (0.89–0.98) 0.94 (0.87–0.97) 14.5 (1.8–115.0)† 0.5 (0.1–3.5) 0.0 0.33 (0.21–0.47) 0.66 (0.49–0.79) 0.36 (0.23–0.52) 0.59 (0.49–0.68) 0.70 (0.61–0.78) 0.61 (0.52–0.69) 0.8 (0.5–1.2) 2.2 (1.6–3.1)† 0.9 (0.6–1.5) 0.51 (0.37–0.64) 0.17 (0.08–0.33) 0.14 (0.06–0.28) 0.86 (0.78–0.92) 0.71 (0.62–0.79) 0.70 (0.61–0.77) 3.6 (2.2–6.1)† 0.6 (0.3–1.2) 0.4 (0.2–1.0) Any RL (n ⫽ 87) SpA DA Malignancy ⱖ3 RLs (n ⫽ 46) SpA DA Malignancy ⱖ3 RLs and age ⱕ50 years (n ⫽ 21) SpA DA Malignancy Grade 3/severe RLs (n ⫽ 22) SpA DA Malignancy Grade 3/severe RLs and age ⱕ50 years (n ⫽ 6) SpA DA Malignancy Any posterior element lesion (n ⫽ 31) SpA DA Malignancy Grade 3 posterior element lesion (n ⫽ 23) SpA DA Malignancy Grade 1 or 2 posterior element lesions (n ⫽ 8) SpA DA Malignancy Any end-plate lesion (n ⫽ 65) SpA DA Malignancy Age ⱕ50 years and any RL or any posterior element lesion (n ⫽ 44) SpA DA Malignancy Positive LR (range) Infinite† 0.0 0.0 * MRI ⫽ magnetic resonance imaging; 95% CI ⫽ 95% confidence interval. † Significant difference (P ⬍ 0.05) versus the other diagnoses. Sacroiliac joint imaging. SI joint scans were available for 89 subjects (48% of the total cohort). Evaluation of the SI joints was included to determine whether its inclusion in a spinal protocol aided in the diagnosis. Fifty-six (63%) of the subjects with SI joint scans were given a diagnosis of SpA and 11 (12%) had normal findings, while the remaining SI joint scans resulted in a diagnosis of degenerative disease in 9 subjects (10%), metastatic disease in 5 (5.6%), and other diagnoses, including abscess and fracture, in 8 (8.9%). In the SpA cohort, all of whom had clinically active disease, 24% of patients had normal findings on the SI joint scans. In 9 (10%) of the patients who had both spinal and SI joint scans, the diagnosis of SpA was correctly established on the basis of the SI joint scan when findings on the spinal scans alone were equivocal. In a further 9 patients (10%), the SI joint scan improved the certainty of a diagnosis of SpA. High-signal STIR lesions in the SI joints of patients with malignancy were often different from those in the SI joints of patients with SpA, with the SI joint scans in those with malignancy having a more heterogeneous, “peppered” appearance. Patients with DA had less frequent and less severe BME lesions on STIR-sequence images of the SI joints than did SpA patients, which supports the findings from previous 1336 Figure 3. Classic appearances of the spine on magnetic resonance imaging (MRI) STIR sequences. A, A degenerative thoracolumbar spine, showing widespread dehydrated extruding intervertebral discs with marked end-plate edema at L2–L3 (short arrow) and a Romanus lesion (RL) anterior-superior to T12 (long arrow). B, A spine with spondylarthritis, showing multiple RLs at T3–T5 (arrows) with wellpreserved intervertebral discs. C, Spinal metastases, showing marked bone destruction and expansile mass at T12 (short arrow), diffuse vertebral body edema at T9 (arrowhead), an RL anterior-superior to T8 (long arrow), and widespread patchy edema at L1–L5. reports (38,39). Two control subjects with no back pain each had a mild lesion evident in the SI joints. Analysis of individual lesions. To explore the global diagnosis findings in more detail, we undertook analyses of different lesions. Romanus lesions. RLs were the most frequently observed lesion (n ⫽ 297 RLs) and were most commonly found in the SpA group (67%, compared with 63% in the DA group and 39% in the malignancy group). However, the diagnostic utility of RLs (all grades of severity included) for SpA was low and not significantly better than for DA or spinal malignancy (Table 1). Other features of malignancy, such as diffuse patchy uptake with low signal on T1 and medium/high signals on STIR, bone destruction, expansile/lytic/sclerotic lesions, or soft-tissue masses, generally allowed a correct diagnosis of spinal malignancy to be made. However, correctly diagnosing SpA while differentiating it from DA was more difficult, because RLs also occurred frequently in DA (Figure 3A) and were not necessarily related to obvious degenerative discs (Figure 2B). BENNETT ET AL The diagnostic utility of RLs for SpA, however, improved with increasing severity of the RLs, with increasing total number of RLs, and with younger age (Table 1). When ⱖ3 RLs (any grade) was considered the minimum cutoff value, the specificity and LR for SpA increased to 0.81 and 2.5, respectively (Table 1). For patients ⱕ50 years old with ⱖ3 (any grade) RLs, the specificity and LR for SpA further increased significantly (P ⬍ 0.05 versus the other diagnoses), to 0.97 and 12.4, respectively. When only severe (grade 3) RLs in patients ⱕ50 years were included, the specificity for SpA increased further still, to 1.00, with an infinite positive LR (P ⬍ 0.05 versus the other diagnoses). Both presence of ⱖ3 (any grade) RLs and severe RLs in those ages ⱕ50 years had a large LR for SpA (LR ⬎10) (Table 1), which represents substantial, and often conclusive, evidence of this disease (40) and indicates that the presence of these lesions in younger patients gives great accuracy to the diagnosis of SpA. Interestingly, the specificity of RLs for SpA was not significantly improved in the absence of degenerative discs (in those with any RLs, sensitivity for SpA 0.68 [range 0.49–0.83], specificity for SpA 0.62 [range 0.50– 0.73]). Moreover, the presence of degenerative discs did not improve the specificity of RLs for degenerative spinal disease (in those with any RLs, sensitivity for DA 0.67 [range 0.45–0.84], specificity for DA 0.45 [range 0.30–0.61]), suggesting that severe RLs are a specific feature of SpA regardless of the presence of degenerative disc disease. Nevertheless, RLs in DA were frequently, but not always, seen adjacent to classic areas of degenerative disc disease in the lower cervical and lumbar spine. Posterior element and spinous process lesions. Posterior element BME lesions were present in 29% of SpA patients, 20% of those with spinal malignancy, and 12% of DA patients. Interestingly, these posterior element lesions were almost exclusively severe (grade 3) in spinal malignancy, but a more even distribution, between grade 1 and grade 3, was observed in SpA and DA. Any grade of posterior element lesion had a high specificity for SpA (0.87), although this was not significantly different from the specificity for DA (0.80) or spinal malignancy (0.82). However, if only mild (grade 1) or moderate (grade 2) lesions were present, the specificity of posterior element lesions for SpA increased (specificity 0.99, LR 14.5), and mild/moderate lesions were significantly more specific for SpA than for DA or spinal malignancy (P ⬍ 0.05). Severe posterior element lesions in patients with malignancy were often associated with other char- SPINAL MRI DIAGNOSIS OF AXIAL SpA Figure 4. Posterior element lesions on magnetic resonance imaging (MRI) STIR sequences from patients with spondylarthritis (SpA) and patients with malignancy. A, Spine of a patient with ankylosing spondylitis (AS), showing a posterior element/costovertebral grade 2 lesion at T12 (arrow). B, Spine of a patient with AS, showing a grade 1 lesion in the superior facet of T6 (short arrow) and grade 1 end-plate lesion (long arrow) in the inferior end plate at T3. C, Spine of a patient with metastatic breast cancer, showing posterior element grade 3 lesions at T9 and T10 (long arrow) and associated grade 3 diffuse vertebral body edema at T9–T10 and bone destruction at T12 (short arrow). acteristic lesions of malignancy, such as bone destruction, as seen in Figure 4C. Of course, the age of the patient is known during routine interpretation of MRI, unlike the blinded procedures for this study. Thus, inclusion of age ⱕ50 years in combination with any RL or any posterior element lesion yielded a significantly greater LR for SpA (LR 3.6) than for DA (LR 0.58; P ⬍ 0.05) or spinal malignancy (LR 0.40; P ⬍ 0.05) (Table 1). Spinous process involvement was seen more frequently in malignancy (27%) and SpA (22%) as compared with DA (10%), but only small, nonsignificant differences in sensitivity and specificity were seen for spinous process BME lesions. Diffuse vertebral body edema. Diffuse vertebral BME was seen in patients with SpA (11%) and in those with degenerative discs (22%) and malignant spines (55%). Diffuse vertebral body edema (any grade/any severity) had a significantly (P ⬍ 0.05) greater specificity for spinal malignancy (specificity 0.84 [range 0.76–0.90], LR 3.44) than for SpA (specificity 0.67 [range 0.57– 0.75], LR 0.33). The specificity and LR of diffuse body edema for spinal malignancy increased further with severe (grade 3) edema only (specificity 0.91 [range 0.84–0.95], LR 4.5; P ⬍ 0.05 versus the other diagnoses). 1337 End-plate edema. End-plate edema was common in the 3 major diagnoses, but more so in DA (66%) than in SpA (33%) or in spinal malignancy (36%). The LR of end-plate edema for DA (LR 2.2) was significantly greater than that for SpA (LR 0.8; P ⬍ 0.05) or that for malignancy (LR 0.9; P ⬍ 0.05) (Table 1). End-plate edema in DA was most frequently seen in the lumbar spine, followed by the cervical spine and thoracic spine (62%, 24%, and 14%, respectively). The lesions were often (63% of end-plate lesions), but not always, seen at classic degenerative sites (L3–L4, L4–L5, L5–S1, and C4–C5, C5–C6, and C6–C7), with L5–S1 accounting for 20% of all end-plate lesions. Location of lesions in SpA. Typical high-signal, STIR-sequence BME (inflammatory) lesions in either the spine or SI joints were seen in 61 (95%) of the 64 SpA patients. The remaining 3 SpA patients with active disease (5%) had normal MRI findings on whole-spine and SI joint scans. Acute SI joint lesions were present in 62% of SpA patients (a further 14% had chronic erosive or ankylosed SI joints, without acute lesions), and 82% of SpA patients had acute spinal lesions. The spinal lesions were located in the T-spine in 62%, L-spine in 59%, and C-spine in 21% of SpA patients. Despite having clinically active disease, 24% of SpA patients had normal findings on the SI joint scans and 18% had no acute lesions on the spinal scans. Conversely, 75% of DA patients (6 of 8) had normal SI joints, but only 9% (4 of 45) had no BME spinal lesions. In patients with normal SI joints but abnormal findings on the spine scans, acute spinal lesions were seen more commonly in those with DA (mean 5.3 lesions per patient) compared with those with SpA (mean 3.3 lesions per patient). In contrast, RLs were more frequent in SpA (61%) than in DA (47%), and severe (grade 3) lesions were more common in SpA (16%) compared with DA (0%). Importantly, in the SpA cohort, 10% of patients had lesions in the cervical or thoracic spine without associated lesions in the lumbar spine of the SI joints. DISCUSSION This is the first study to systematically investigate the diagnostic utility of characteristic SpA lesions on MRI (5,41) in a cohort of patients with back pain with mixed pathologic features. We showed that MRI assessment alone, without any clinical details, had good diagnostic utility in 3 causes of back pain, SpA, DA, and malignancy, with 72% accuracy for all 3 diagnoses and 75% accuracy for SpA. When the MRI lesions were 1338 BENNETT ET AL Figure 5. Characteristic spondylarthritis (SpA) lesions in patients with SpA and patients with malignancy. A, Presence of diffuse vertebral body edema (arrow) and a Romanus lesion (RL) (arrowhead) in selected thoracic vertebrae, both consistent with a diagnosis of SpA. B, Extended spinal views of the same patient as in A, showing further features of SpA (spinous process edema [arrow]). C, Presence of an RL (arrow) and diffuse vertebral body edema (arrowhead), consistent with a diagnosis of SpA, in selected thoracic vertebrae of a patient. D, Extended spinal views of the same patient as in C, showing obvious features of malignant disease with bone destruction and a soft-tissue mass at T12 (arrow). interpreted in clinical context, especially according to the patient’s age, and in detail, they were found to be highly diagnostic of SpA. These findings could help in establishing a definite diagnosis of axial SpA in the 23% of cases in which the SI joints are spared (13). However, the study also shows that typical SpA lesions are in fact also common in other causes of back pain. In older patients with later disease presentations, particularly in those in whom only a few MRI lesions are present, the findings indicate that considerable caution is needed when ascribing a diagnosis. We found that RLs are common in SpA, but also are common in degenerative and malignant disease. However, our detailed analysis of RLs indicated that the presence of numerous or severe RLs in subjects ages ⱕ50 years has high diagnostic accuracy for SpA. Although patients with malignant spinal disease also often had RLs present and could be misdiagnosed as having SpA at a vertebral level, this was in fact a very uncommon misdiagnosis at the patient level, since the overall MRI picture was clearly indicative of malignant disease, due to spinal imaging features elsewhere (Figure 5). We have previously noted that the MRI pattern of disease in the DIP joints in patients with PsA and those with OA is often similar (14). Recent findings from a non–fat- suppressed MRI study also suggested some potential overlap in chronic changes between SpA and spinal OA (42). Therefore, the presence of RLs on MRI in patients with DA indicates that MRI findings of such lesions in SpA must be carefully interpreted. This study shows that although posterior element lesions on high-signal STIR images were less frequent than RLs in SpA, mild and moderate posterior element lesions were highly specific, having an extremely high LR for a diagnosis of SpA. In contrast, severe posterior element lesions were suggestive of malignancy, especially in older patients, and other MRI and clinical signs of malignancy should be sought and excluded before a diagnosis of SpA is entertained. This study was not designed to develop a new spinal scoring system but to describe the frequency, location, and morphology of spinal lesions seen in SpA and other diseases, as occurs in clinical practice, and to investigate the diagnostic utility of these lesions. We believe that the high specificity of posterior element lesions for SpA implies that the addition of posterior element assessment in existing scoring systems might further improve their utility. We found that degenerative discs on MRI were of low diagnostic utility when trying to distinguish between back pain in SpA and that in DA. It has SPINAL MRI DIAGNOSIS OF AXIAL SpA previously been accepted that RLs or end-plate lesions occurring on either side of degenerative discs are likely to be degenerative in nature (31,35). However, in this study, degenerative discs were common in normal subjects, SpA patients, and patients with spinal malignancy, as well as in patients with DA. We recognize that the present findings come from a cohort of SpA patients whose mean age was 41 years and whose mean disease duration was 8.5 years, and that the specificity of degenerative discs for DA would probably have been higher if the studied subjects were under age 30 years. However, we specifically aimed to study a wide age range to capture all the possible ages at onset of SpA (16–18). In particular, we were interested in studying the 30– 50-year-old age group, since we believe that these subjects may provide the most difficult diagnostic imaging dilemmas because of the common acute inflammatorylike MRI lesions that occur in degenerative spines in this age group (19) and which may be mistaken for SpA. Our findings confirmed that end-plate lesions had a significantly higher LR for a diagnosis of DA. However, this finding should be interpreted with caution, since endplate lesions were also common in SpA and malignancy. The high diagnostic utility of the above-described RLs and posterior element lesions for SpA could be of considerable diagnostic importance, since chronic back pain is often investigated by nonrheumatologists and SI joint scanning may be excluded. Another important consideration is which spinal regions to scan. In most cases of suspected SpA, lumbar spine and SI joint MRI sequences are clinically investigated. However, acute BME lesions in SpA were most frequently seen in the thoracic spine (62%), which is consistent with previously reported data in patients with AS (43). In addition, 10% of SpA patients in the present cohort had only thoracic or cervical spine lesions, and 24% had normal findings on their SI joint scans. This indicates that scanning the whole spine as well as the SI joints may be a more diagnostically useful tool when investigating potential SpA. Also of interest is that 5% of SpA patients had no acute inflammatory lesions on MRI STIR sequences or on scans of the whole spine and SI joints, despite having clinically active disease. This study has some limitations. Only sagittal images of the spine were scored; therefore, some lesions that may be evident only on coronal planes may have been missed. Also, MRI findings were evaluated and the most probable single diagnosis was given based on these findings; therefore, the possibility of coexisting conditions was not taken into account. Finally, the present SpA cohort generally had established disease, with only 1339 a minority having early disease. Nevertheless, the evidence suggests that the MRI appearances of early SpA are similar to those of late disease (3,38), and studying a cohort with a mean disease duration of 8.5 years may have captured a time point early enough for up to 40% of subjects to not have developed diagnostic radiographic features (44). In this scenario, specific MRI lesions would be of great assistance in making diagnoses. Our study has thus shown that MRI has good diagnostic utility in SpA, with a 75% rate of correct diagnosis made by scan alone. However, characteristic SpA lesions on MRI are also frequently present in those with malignant, and particularly degenerative, spinal disease. This may lead to misdiagnosis, especially in older people, and therefore the findings need to be examined in closer detail. Severe or multiple RLs on MRI in younger patients and mild/moderate posterior element lesions are highly diagnostic of SpA, irrespective of any other clinical or laboratory parameter. AUTHOR CONTRIBUTIONS All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Bennett and Prof. McGonagle had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study conception and design. Bennett, Rehman, Hensor, MarzoOrtega, McGonagle. Acquisition of data. Bennett, Rehman, Hensor, Marzo-Ortega, McGonagle. Analysis and interpretation of data. Bennett, Rehman, Hensor, Marzo-Ortega, McGonagle. REFERENCES 1. Mau W, Zeidler H, Mau R, Majewski A, Freyschmidt J, Stangel W, et al. Clinical features and prognosis of patients with possible ankylosing spondylitis: results of a 10-year follow-up. J Rheumatol 1988;15:1109–14. 2. Puhakka KB, Jurik AG, Schiottz-Christensen B, Hansen GV, Egund N, Christiansen JV, et al. MRI abnormalities of sacroiliac joints in early spondylarthropathy: a 1-year follow-up study. Scand J Rheumatol 2004;33:332–8. 3. Bennett AN, McGonagle D, O’Connor P, Hensor EM, Sivera F, Coates LC, et al. Severity of baseline magnetic resonance imaging–evident sacroiliitis and HLA–B27 status in early inflammatory back pain predict radiographically evident ankylosing spondylitis at eight years. Arthritis Rheum 2008;58:3413–8. 4. Benjamin M, McGonagle D. The anatomical basis for disease localisation in seronegative spondyloarthropathy at entheses and related sites. J Anat 2001;199(Pt 5):503–26. 5. Maksymowych WP, Lambert RG. Magnetic resonance imaging for spondyloarthritis: avoiding the minefield. J Rheumatol 2007;34: 259–65. 6. Marzo-Ortega H, McGonagle D, O’Connor P, Emery P. Efficacy of etanercept in the treatment of the entheseal pathology in resistant spondylarthropathy: a clinical and magnetic resonance imaging study. Arthritis Rheum 2001;44:2112–7. 7. Marzo H, McGonagle D, O’Connor P, Green MJ, Pease C, 1340 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. Wakefield R, et al. Fat suppressed MRI in spinal disease in early spondyloarthropathy [abstract]. Arthritis Rheum 1998;41 Suppl 9:S355. Rudwaleit M, van der Heijde D, Khan MA, Braun J, Sieper J. How to diagnose axial spondyloarthritis early. Ann Rheum Dis 2004;63: 535–43. Maksymowych WP, Inman RD, Salonen D, Dhillon SS, Krishnananthan R, Stone M, et al. Spondyloarthritis Research Consortium of Canada magnetic resonance imaging index for assessment of spinal inflammation in ankylosing spondylitis. Arthritis Rheum 2005;53:502–9. Braun J, Baraliakos X, Golder W, Brandt J, Rudwaleit M, Listing J, et al. Magnetic resonance imaging examinations of the spine in patients with ankylosing spondylitis, before and after successful therapy with infliximab: evaluation of a new scoring system. Arthritis Rheum 2003;48:1126–36. Rudwaleit M, Schwarzlose S, Listing J, Brandt J, Braun J, Sieper J. Is there a place for magnetic resonance imaging (MRI) in predicting a major clinical response (BASDAI 50) to TNF alpha blockers in ankylosing spondylitis [abstract]. Arthritis Rheum 2004;50 Suppl 9:S211. Marzo-Ortega H, McGonagle D, Jarrett S, Haugeberg G, Hensor E, O’Connor P, et al. Infliximab in combination with methotrexate in active ankylosing spondylitis: a clinical and imaging study. Ann Rheum Dis 2005;64:1568–75. Rudwaleit M, Schwarzlose S, Hilgert ES, Listing J, Braun J, Sieper J. MRI in predicting a major clinical response to anti-tumour necrosis factor treatment in ankylosing spondylitis. Ann Rheum Dis 2008;67:1276–81. Tan AL, Grainger AJ, Tanner SF, Emery P, McGonagle D. A high-resolution magnetic resonance imaging study of distal interphalangeal joint arthropathy in psoriatic arthritis and osteoarthritis: are they the same? Arthritis Rheum 2006;54:1328–33. Dougados M, van der Linden S, Juhlin R, Huitfeldt B, Amor B, Calin A, et al, and the European Spondylarthropathy Study Group. The European Spondylarthropathy Study Group preliminary criteria for the classification of spondylarthropathy. Arthritis Rheum 1991;34:1218–27. Olivieri I, Salvarani C, Cantini F, Ciancio G, Padula A. Ankylosing spondylitis and undifferentiated spondyloarthropathies: a clinical review and description of a disease subset with older age at onset. Curr Opin Rheumatol 2001;13:280–4. Shankaranarayana S, Tan AL, Madden J, McGonagle D. Fat-suppression magnetic resonance imaging in the diagnosis of late-onset axial spondyloarthropathy. Ann Rheum Dis 2007;66: 1405–6. Toussirot E, Wendling D. Late-onset ankylosing spondylitis and related spondylarthropathies: clinical and radiological characteristics and pharmacological treatment options. Drugs Aging 2005; 22:451–69. Kuisma M, Karppinen J, Niinimaki J, Ojala R, Haapea M, Heliovaara M, et al. Modic changes in endplates of lumbar vertebral bodies: prevalence and association with low back and sciatic pain among middle-aged male workers. Spine 2007;32: 1116–22. Van der Heijde DM, Landewe RB, Hermann KG, Jurik AG, Maksymowych WP, Rudwaleit M, et al. Application of the OMERACT filter to scoring methods for magnetic resonance imaging of the sacroiliac joints and the spine: recommendations for a research agenda at OMERACT 7. J Rheumatol 2005;32: 2042–7. Carley S, Dosman S, Jones SR, Harrison M. Simple nomograms to calculate sample size in diagnostic studies. Emerg Med J 2005; 22:180–1. Hermann KG, Landewe RB, Braun J, van der Heijde DM. Magnetic resonance imaging of inflammatory lesions in the spine BENNETT ET AL 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. in ankylosing spondylitis clinical trials: is paramagnetic contrast medium necessary? J Rheumatol 2005;32:2056–60. Ostergaard M, Edmonds J, McQueen F, Peterfy C, Lassere M, Ejbjerg B, et al. An introduction to the EULAR–OMERACT rheumatoid arthritis MRI reference image atlas. Ann Rheum Dis 2005;64 Suppl I:i3–7. Landewe RB, Hermann KG, van der Heijde DM, Baraliakos X, Jurik AG, Lambert RG, et al. Scoring sacroiliac joints by magnetic resonance imaging: a multiple-reader reliability experiment. J Rheumatol 2005;32:2050–5. Maksymowych WP, Dhillon SS, Park R, Salonen D, Inman RD, Lambert RG. Validation of the Spondyloarthritis Research Consortium of Canada magnetic resonance imaging spinal inflammation index: is it necessary to score the entire spine? Arthritis Rheum 2007;57:501–7. Crowther SM, Lambert RG, Dhillon SS, Maksymowych WP. High frequency of inflammatory lesions in the posterior structures of the spine in patients with ankylosing spondylitis (AS): a systematic evaluation by MRI [abstract]. Arthritis Rheum 2006;54 Suppl 9:S793. Bochkova A, Lewschakova A, Bunchuk N. Inflammatory spinal MRI lesions in patients with ankylosing spondylitis and short history of axial pain prevail in the posterior structures. Ann Rheum Dis 2008;67 Supp II:379. Lukas C, Braun J, van der Heijde D, Hermann KG, Rudwaleit M, Ostergaard M, et al. Scoring inflammatory activity of the spine by magnetic resonance imaging in ankylosing spondylitis: a multireader experiment. J Rheumatol 2007;34:862–70. Van der Heijde D, Landewe R, Hermann KG, Rudwaleit M, Ostergaard M, Oostveen A, et al. Is there a preferred method for scoring activity of the spine by magnetic resonance imaging in ankylosing spondylitis? J Rheumatol 2007;34:871–3. MacVicar D. Imaging of the spine in patients with malignancy. Cancer Imaging 2006;6:S22–6. Rankine JJ, Hutchinson CE, Hughes DG. MRI of lumbar spondylosis: a comparison of sagittal T2 weighted and three sequence examinations. Br J Radiol 1997;70:1112–21. McGonagle D, Gibbon W, O’Connor P, Green M, Pease C, Emery P. Characteristic magnetic resonance imaging entheseal changes of knee synovitis in spondylarthropathy. Arthritis Rheum 1998;41: 694–700. McGonagle D, Marzo-Ortega H, O’Connor P, Gibbon W, Pease C, Reece R, et al. The role of biomechanical factors and HLA–B27 in magnetic resonance imaging–determined bone changes in plantar fascia enthesopathy. Arthritis Rheum 2002;46: 489–93. Haibel H, Rudwaleit M, Brandt HC, Grozdanovic Z, Listing J, Kupper H, et al. Adalimumab reduces spinal symptoms in active ankylosing spondylitis: clinical and magnetic resonance imaging results of a fifty-two–week open-label trial. Arthritis Rheum 2006;54:678–81. Karchevsky M, Schweitzer ME, Carrino JA, Zoga A, Montgomery D, Parker L. Reactive endplate marrow changes: a systematic morphologic and epidemiologic evaluation. Skeletal Radiol 2005; 34:125–9. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159–74. McGee S. Simplifying likelihood ratios. J Gen Intern Med 2002; 17:646–9. Marzo-Ortega H, McGonagle D, O’Connor P, Hensor EM, Bennett AN, Green M, et al. Baseline and one year magnetic resonance imaging of the sacroiliac joints and lumbar spine in early inflammatory back pain: the relationship between HLA-B27, symptoms and disease extent and persistence. Ann Rheum Dis 2008;67 Sup II:ii377–8. Rudwaleit M, Vahldiek J, Wenz J, Brandt H, Appel H, Haibel H, et al. Sensitivity and specificity of magnetic resonance imaging SPINAL MRI DIAGNOSIS OF AXIAL SpA (MRI) of the sacroiliac joints in patients with suspected early ankylosing spondylitis [abstract]. Arthritis Rheum 2006;54 Suppl 9:S793–4. 40. Jaeschke R, Guyatt GH, Sackett DL, the Evidence-Based Medicine Working Group. Users’ guides to the medical literature. III. How to use an article about a diagnostic test. B. What are the results and will they help me in caring for my patients? JAMA 1994;271:703–7. 41. Zochling J, Baraliakos X, Hermann KG, Braun J. Magnetic resonance imaging in ankylosing spondylitis. Curr Opin Rheumatol 2007;19:346–52. 1341 42. Kim NR, Choi JY, Hong SH, Jun WS, Lee JW, Choi JA, et al. “MR corner sign”: value for predicting presence of ankylosing spondylitis. AJR Am J Roentgenol 2008;191:124–8. 43. Baraliakos X, Landewe R, Hermann KG, Listing J, Golder W, Brandt J, et al. Inflammation in ankylosing spondylitis: a systematic description of the extent and frequency of acute spinal changes using magnetic resonance imaging. Ann Rheum Dis 2005;64: 730–4. 44. Rudwaleit M, Khan MA, Sieper J. The challenge of diagnosis and classification in early ankylosing spondylitis: do we need new criteria? Arthritis Rheum 2005;52:1000–8. DOI 10.1002/art.24455 Clinical Images: Cement embolism following percutaneous vertebroplasty The patient, a 78-year-old woman, presented with back pain, which she had been experiencing for several months. Plain radiography revealed a compression fracture of the first and second vertebral bodies and radiodense materials in the first vertebral body (left). In addition, along the right border of the lumbar vertebrae was a radiopaque vascular structure, which seemed to be the ascending lumbar vein (left). A radiograph of the chest showed branching, tubular, radiodense lines in the lungs (right). After further inquiry, the patient revealed that she had undergone percutaneous vertebroplasty for an osteoporotic lumbar compression fracture 3 years prior. Cement embolism due to the leakage of acrylic bone cement was diagnosed. Hyeok-Jae Ko, MD Kyung-Su Park, MD Chul-Soo Cho, MD Ho-Youn Kim, MD The Catholic University of Korea College of Medicine Seoul, Korea
1/--страниц