European Journal of Radiology 83 (2014) 366–370
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Comparison of apparent diffusion coefficient in spondylarthritis axial active inflammatory lesions and type 1 modic changes Benjamin Dallaudière a,b,c,∗ , Raphaël Dautry a,c,1 , Pierre-Marie Preux d,2 , Anne Perozziello c,e,3 , Julien Lincot a,1 , Elisabeth Schouman-Claeys a,c,1 , Jean-Michel Serfaty a,b,c,1 a
Hôpital Bichat – Claude Bernard, Service de Radiologie, 46, rue Henri Huchard, Paris 75018, France Hôpital Bichat – Claude Bernard, Inserm U698, 46, rue Henri Huchard, Paris 75018, France c Université Paris Diderot, Paris, France d Faculté de Médecine de Limoges, Unité fonctionelle de recherche clinique et de biostatistique, hôpital Le Cluzeau, 23, avenue Dominique Larrey, 87042 Limges Cedex, France e Hôpital Bichat – Claude Bernard, Unité de recherche clinique, 46, rue Henri Huchard, Paris 75018, France b
a r t i c l e
i n f o
Article history: Received 31 January 2013 Received in revised form 26 August 2013 Accepted 12 October 2013 Keywords: Spondylarthritis Spondylitis Diffusion weighted imaging Apparent diffusion coefficient Modic change
a b s t r a c t Objective: The goal of this study was to evaluate whether the values of ADC in spondylarthritis axial active inflammatory lesions are different from ADC values in type 1 Modic changes. Subjects and methods: 95 patients with recent lumbar pain, including 46 patients with diagnosed or suspected spondylarthritis and 49 patients with purely degenerative history, underwent spine MRI. T1w, STIR, and diffusion-weighted images (DWI) were obtained. Two musculoskeletal radiologists interpreted the images. Axial active inflammatory lesions from the SpA group and type 1 Modic changes from the degenerative group were identified on T1w and STIR sequences. ADC values from these lesions and from healthy subchondral bone were compared. Results: All axial active inflammatory lesions (n = 27) and type 1 Modic changes (n = 22) identified in T1w and STIR images were visible on DWI. ADC values were significantly higher (p < 0.05) for axial active inflammatory lesions (median = 0.788 × 10−3 mm2 /s, IQR 25–75 [0.7 × 10−3 mm2 /s; 0.9 × 10−3 mm2 /s]) than for type 1 Modic changes (median = 0.585 × 10−3 mm2 /s, IQR 25–75 [0.55 × 10−3 mm2 /s; 0.60 × 10−3 mm2 /s]) and normal subchondral bone (median = 0.443 × 10−3 mm2 /s, IQR 25–75 [0.40 × 10−3 mm2 /s; 0.50 × 10−3 mm2 /s]). Intra-class correlation coefficients for intra- and inter-reader ADC values comparison were excellent (0.89 and 0.98 respectively). Conclusion: DWI is a sensitive and fast sequence that offer the possibility of quantifying diffusion coefficients of the lesions, which could help to discriminate between spondylarthritis axial active inflammatory and type 1 Modic changes. © 2013 Elsevier Ireland Ltd. All rights reserved.
1. Introduction
Abbreviations: ADC, apparent diffusion coefficient; DWI, diffusion weighted imaging; ROI, region of interest; SpA, spondylarthritis; T1w, T1 weighted; ASAS/OMERACT, Assessment of Spondylarthritis International Society/Outcome Measures in Rheumatoid Arthritis Clinical Trials; DCE-MRI, dynamic contrastenhanced MRI. ∗ Corresponding author at: Hôpital Bichat – Claude Bernard, Service de Radiologie, 46, rue Henri Huchard, Paris 75018, France. Tel.: +33 1 40 25 80 80; fax: +33 1 40 25 88 73. E-mail addresses: [email protected], [email protected] (B. Dallaudière), raphael [email protected] (R. Dautry), [email protected] (P.-M. Preux), [email protected] (A. Perozziello), [email protected] (J. Lincot), [email protected] (E. Schouman-Claeys), [email protected] (J.-M. Serfaty). 1 Tel.: +33 1 40 25 80 80; fax: +33 1 40 25 88 73. 2 Tel.: +33 5 55 05 66 29. 3 Tel.: + 33 1 40 25 79 31. 0720-048X/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ejrad.2013.10.009
Lumbar pain is one of the most frequently reported symptoms, with a high impact on life quality [1]. When there is no evident traumatic, infectious or neoplastic context, back pain usually is of rheumatismal or degenerative cause. Most times clinical information in addition to imaging studies are sufficient to differentiate these two patterns. However the differential diagnosis can be tricky when clinical dates are scarce, and when lesions do not display a typical imaging aspect. Inflammatory spondylarthritis (SpA) are common diseases, whose frequency is estimated to be between 0.3 and 1.9% in the general population [1,2]. Because of the young age of the SpA patients, acute lumbar pains are usually attributed to SpA axial active inflammatory lesions but degenerative lesions such as type 1 Modic changes are a frequent cause for back pain and may be overlooked
B. Dallaudière et al. / European Journal of Radiology 83 (2014) 366–370
in young patients. Such misdiagnosis may lead to overtreatment of patients with either 3 months corticotherapy or immunosuppressive therapy. MRI, with T1 weighted and STIR sequences, is widely used to assess back pain patients. Differential diagnosis between axial active inflammatory lesions and Modic1 lesions can be tricky as lesion signal does not permit to discriminate between these two types of acute lesions [3,4], and anatomical topography sometimes overlap [5]. Recent works have evaluated new MRI technics for assessement and follow-up of spondylarthritis inflammatory lesion. Dynamic contrast-enhanced MRI (DCE-MRI) has been reported to be valuable in detecting early- and late-stage inflammation in the sacroiliac joints of patients with spondylarthropathy, as well as for the therapeutic follow-up of SpA [6,7]. Although in this specific clinical setting this technique has shown promising results, it is not routinely used in imaging of SpA. Another recent study has shown that diffusion weighted MRI (DWI) might be effective to differentiate acute degenerative lesions from early inflammatory sacro-iliitis [8]. Without the need of contrast injection, this makes DWI a potentially effective tool to differentiate axial active inflammatory lesions and type 1 Modic changes. To our knowledge, the role of diffusion weighted MRI with ADC in this application has not been studied [9,10]. The goal of our study was therefore to compare ADC values of axial active inflammatory lesions in spondylarthritis patients to ADC values of type 1 Modic changes in patients with a purely degenerative disease.
2. Materials and methods 2.1. Patients We conducted a mono-centric observational cohort study on 46 consecutive spondylarthritis patients and 49 purely degenerative patients from November 2010 to June 2012. For the spondylarthritis group, inclusion criteria were probable SpA (Amor score = 5) or confirmed SpA (Amor score ≥ 6) [11] presenting with a recent back pain (duration of less than a month). For the degenerative group, inclusion criteria were recent back pain with Amor score < 5 and no antecedent of personal of familial rheumatismal inflammatory disease. All patients were referred to our imaging department by the Rheumatology and Internal Medicine departments for an MRI of the spine. Exclusion criteria included contraindication to MRI (pregnancy, metallic implants, and claustrophobia) and suspicion of spine infection. Fifteen patients of the spondylarthritis group belonged to the DESIR study [12]. All patients were informed of the study procedure and gave their informed consent.
367
0 and 1000 s/mm2 (we chose this high b-value to enhance DWI sensitivity [8]). On the thoraco-lumbar spine, DWI parameters were: sagittal slices, TR 6000 ms, TE 1114 ms number of excitations, 2; duration 2.06 min, slice thickness, 5 mm; intersection gap, none. On the cervico-thoracic spine, DWI parameters were: axial slices, TR 7075 ms, TE 87.4 ms number of excitations, 2; duration 4.57 min, slice thickness, 5 mm; intersection gap, none. Axial slices were used to benefit from a low acquisition time and a lesser sensitivity to motion artifacts, as electrocardiogram (EKG) gated sagittal DWI are longer to acquire.
2.3. Image interpretation Patients’ identities were removed from all images. Two musculoskeletal radiologists blindly assessed images, in random patient order, for the presence of axial active inflammatory lesions and type 1 Modic changes. T1 and STIR weighted images were looked upon first, followed immediately by DWI analysis with ADC measurements on lesions identified on T1 and STIR images. Both readers, according to the same protocol, performed a second interpretation two weeks later. The presence of bone marrow edema (T1 hypo signal and STIR hyper signal) was considered as a marker of acute disease. ADC maps were obtained from DWI data, on a GE Healthcare Advantage Windows, 4.2 Workstation. In order to be more reproducible, measurements were made on the Kodac PACS system® using circular ROI having an area of at least 4 pixels, or 4 mm × 4 mm (information given by the PACS system) that was positioned on active inflammatory lesions and type 1 Modic changes, as well as on 2 different locations on normal subchondral bone.
2.4. Statistical methods Statistical analysis was performed using the SAS® software. ADC values for active inflammatory lesions, type 1 Modic changes and healthy subchondral bone were expressed as follows (median and inter quantile range (IQR)) and compared using a Wilcoxon signed rank sum test, for paired values, not normally distributed. ADC cut-off was determined using the boxplot graphic. We considered p < 0.05 as significant. Intra- and inter-reader intraclass correlation coefficients for ADC value measurement were calculated.
3. Results 3.1. Population
2.2. MRI All patients underwent MRI of the either the entire spine (n = 32 in the SpA group and n = 14 in the degenerative group) or the thoraco-lumbar spine (n = 14 in the SpA group and n = 35 in the degenerative group). MRI was performed using a 1.5-Tesla MR scanner (Twin Speed HDX: GE® Healthcare) with a spine coil (CTI Array by USAI; 6 elements, 6 channels). The following sequences were obtained: sagittal fast spin-echo T1-weighted (TR/TE: 660/9.5 ms, duration 2.53 min, FOV 48 cm, 13 slices), and sagittal STIR (TR/TE: 4700/68 ms, duration 3.36 min, FOV 48 cm, 13 slices) sequences. Slice thickness was 3 mm; number of excitations, 2; and intersection gap, none. In addition to these previous sequences, DWI (without suppression of fat signal) was performed using a single-shot spin echo-planar imaging sequence with diffusion gradient (b value) of
Ninety-five patients were included in this study, 46 in the spondylarthritis group (27 women) and 49 in the degenerative group (28 women). In the spondylarthritis group, mean age was 43 years (SD 5.3 years). Twenty-seven patients were clinically suspected of SpA and 19 patients were clinically diagnosed with SpA (including 14 patients with ankylosing spondylitis, 3 with psoriatic arthritis and 2 with spondylarthritis associated to inflammatory bowel disease). Mean back pain duration before MRI was 19 days (±3.2 days). Among the 27 patients with suspected SpA and the 19 patients with SpA, 0 and 15 respectively were receiving medical treatment (either anti-TNF or corticotherapy) at the time of imaging. In the degenerative group, mean age was 52 years (SD 4.1 years), mean back pain duration before MRI was 21 days (±2.8 days), and none had had therapeutic spinal injection.
368
B. Dallaudière et al. / European Journal of Radiology 83 (2014) 366–370
Fig. 1. (a)–(c) T1 (a), STIR (b), DWI (c) and ADC map (d) sagittal slices showing a rheumatismal acute L5-S1 spondylodiscitis.
Fig. 2. (a)–(c) T1 (a), STIR (b), DWI (c) and ADC map (d) sagittal slices showing a L4–L5 Modic 1 lesion.
3.2. Lesion detection by T1 and STIR weighted images In the spondylarthritis group, we found 27 typical axial active inflammatory lesions (17 anterior spondylitis, 2 discitis, 5 posterior arthritis and 3 costo-vertebral or costo-transversary arthritis) in 22 patients. In the degenerative group, we found 22 typical type 1 Modic changes in 14 patients, all located on vertebral endplates (20 on the lumbar spine and 2 on the cervical spine). 3.3. DWI study and ADC measurement All lesions identified on T1w and STIR was visible on DWI. ADC measurements were performed on 17 axial active inflammatory lesions (17 spondylitis, Fig. 1) out of 27 lesions detected, 22 type
1 Modic lesions (Fig. 2), and 92 subchondral bone with no lesion (Fig. 3). We excluded the 10 remaining axial active inflammatory lesions (posterior arthritis, costo-vertebral or costo-transversary arthritis lesions) from ADC analysis due to small size. ADC values were higher for axial active inflammatory lesions (median = 0.788 × 10−3 mm2 /s, IQR 25–75 [0.7 × 10−3 mm2 /s; 0.9 × 10−3 mm2 /s]) than for type 1 Modic changes (median = 0.585 × 10−3 mm2 /s, IQR 25–75 [0.55 × 10−3 mm2 /s; 0.60 × 10−3 mm2 /s]) and for normal subchondral bone (median = 0.445 × 10−3 mm2 /s, IQR 25–75 [0.40 × 10−3 mm2 /s; 0.50 × 10−3 mm2 /s]) (Fig. 4). Median ADC values were significantly different between axial active inflammatory lesions and Md1 (p = 0.0005), between axial active inflammatory lesions and normal subchondral bone (p < 0.0001) and between type 1 Modic changes and normal
Fig. 3. (a)–(c) T1 (a), STIR (a), DWI (b) and ADC map (c) sagittal slices of a normal lumbar spine.
B. Dallaudière et al. / European Journal of Radiology 83 (2014) 366–370
369
Fig. 4. Graphical view of the included patients, the identified lesions and the calculated ADC values.
subchondral bone (p = 0.0037). We used the boxplot graphic to visually determine the best ADC cut-off value to differentiate between axial active inflammatory lesions and type 1 Modic changes in our patients. The best ADC cut-off value was 0.58 × 10−3 mm2 /s (Fig. 5). ADC values were superior to 0.58 × 10−3 mm2 /s in 15 out of 17 inflammatory lesions in which ADC could be measured, and inferior to 0.58 × 10−3 mm2 /s in 19 out of 22 type 1 Modic changes. 3.4. Intra- and inter-reader agreement Inter-reader intraclass correlation coefficients for ADC measurements were very good (ICC = 0.98 and 0.89 respectively). Intra-reader agreement was very good as well, for both radiologists (ICC = 0.98 and 0.92 respectively). 4. Discussion Our results show for the first time that in case of recent back pain, ADC values of axial active inflammatory lesions in spondylarthritis patients are different from ADC values of type 1 Modic changes in patients with purely degenerative lesions. DWI is a fast MRI sequence, which offers without any contrast media injection, an approximation of micro vascular perfusion and diffusion of water molecules in the interstitium. Its role for assessment of various types of spine vertebral and discal lesions has been increasingly studied, in the field of degenerative [13],
Fig. 5. Graphical view of IQR 25–75 for ADC values for axial active inflammatory lesion, type 1 Modic changes and normal subchondral bone.
infectious [14] and neoplastic lesions [15,16]. Recently, Gaspersic et al. [6] have shown a correlation between clinical efficacy of ankylosing spondylitis treatment and decrease of ADC in corresponding inflammatory lesions, adding weight to the assumption that ADC value correlate to clinical intensity of inflammatory process. Bozgeyik et al. [8] have also observed that, in SpA patients, ADC values of sacro-iliitis lesions and subchondral sacro-iliac bone were different. Our results are consistent with these findings and suggest that ADC might be different between type 1 Modic changes and spondylarthritis inflammatory spine lesions. The pathophysiological substratum for the difference of ADC between SpA lesions and Type 1 Modic changes is unclear. In their original study, Modic et al. [17] analyzed histopathological sections of type 1 changes and found that they were associated with disruption and fissuring of endplates and formation of a fibrovascular granulation tissue. They concluded that type 1 changes correspond to the inflammatory stage of disk degenerative disease and indicate an active degenerative process. Concerning SpA lesions, although the underlying causes of inflammation are not well understood, a large body of evidence indicates that the pathology is at least partly mediated by immune cells and that TNF-␣ plays a central role in a majority of pathologic mechanisms that drive inflammation and damage. Several features of inflammatory arthritis, including angiogenesis and bone erosions without fissures, have been attributed to the extracellular matrix degradation by metallo-proteinases with also an increase in the number of osteoclasts present in the SpA bone [18]. The difference of ADC between these two types of lesions could be linked to the underlying mechanisms of inflammation, or could reflect a more severe inflammation in SpA lesions. The difference of ADC between SpA active inflammatory spine changes and type 1 Modic changes could be interesting in patients presenting non-specific bone lesions on MR images. Indeed, this differential diagnosis is a classical pitfall when interpreting inflammatory changes in spondylarthritis patient, and is sometime not feasible on conventional MRI results alone [19,20]. Our study has several limitations. First, in our population, the diagnostic of spondylarthritis was based on the Amor criteria, in order to have two clearly defined populations. However, the Amor criteria are not always fulfilled at the early stages of spondylarthritis, which is why it would be highly beneficial to be able to differentiate spondylarthritis lesions from Modic 1 changes using only MRI sequences. Second, we did not compare DWI/ADC to DCE-MRI. DCE-MRI has shown to better quantify inflammation than STIR sequences in the therapeutic follow-up of SpA [6], and so could be a good technique
370
B. Dallaudière et al. / European Journal of Radiology 83 (2014) 366–370
to differentiate spondylarthritis spine lesions and type 1 Modic changes. Neither did we compare DWI/ADC to the ASAS/OMERACT criteria. Indeed, the main goal of our study was to assess whether ADC values were different in the inflammatory lesions of spondylarthritis and in the type 1 Modic changes of our population. Based on these results as described above, our next step will be therefore to compare the diagnostic accuracy of DWI/ADC to other diagnostic tools such as the ASAS/OMERACT criteria and DCE-MRI. Third, a technical limitation is the low spatial resolution of the DWI sequence (in our study 1 mm × 1 mm with a 5 mm slice thickness) compared to the spatial resolution of T1w and STIR sequences. This explains our decision to exclude lesions smaller than 4 mm for ADC measurements. A solution to this problem could be to use DWI specifically on the lesion to characterize, using thinner slice, smaller FOV and systematic electrocardiogram (EKG) gated sequences. However, despite this limitation, we were able to find significant differences between our groups. Moreover, we found normal subchondral bone values consistent with the literature [21,22]. Future improvement in spatial resolution should help confirm differences in ADC measurement. Regarding the choice of b, we used a high diffusion gradient (b value) of 1000 s/mm2 to enhance the sensitivity of DWI [10] compared to lower b values [8]. Another limitation is related to the use of different axial slice orientation in cervico-thoracic acquisitions instead of the classical sagittal acquisition plan. This axial slice orientation was chosen as sagittal EKG-gated sequences (to avoid cardiac motion artifact) have a long acquisition time. Axial non EKG-gated acquisition almost avoids motion artifacts with a limited cost regarding time acquisition. Finally, the number of patients included in our study was limited and future prospective studies with patient and lesion follow-up will be needed to confirm our initial results. In conclusion, DWI is a sensitive, fast sequence and does not require a contrast agent. Our study suggests that DWI offer the possibility of quantifying diffusion coefficients of the lesions, which helps to discriminate between normal and involved active inflammatory and degenerative subchondral bone. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ejrad.2013. 10.009. References [1] Sieper J, Rudwaleit M, Khan MA, Braun J. Concepts and epidemiology of spondyloarthritis. Best Pract Res Clin Rheumatol 2006;20:401–17.
[2] Saraux A, Guillemin F, Guggenbuhl P, et al. Prevalence of spondyloarthropathies in France: 2001. Ann Rheum Dis 2005;64:1431–5. [3] Althoff CE, Feist E, Burova E, et al. Magnetic resonance imaging of active sacroiliitis: do we really need gadolinium. Eur J Radiol 2009;71:232–6. [4] Jevtic V, Kos-Golja M, Rozman B, McCall I. Marginal erosive discovertebral ‘Romanus’ lesions in ankylosing spondylitis demonstrated by contrast enhanced Gd-DTPA magnetic resonance imaging. Skeletal Radiol 2000;29:27–33. [5] Hermann K-GA, Althoff CE, Schneider U, et al. Spinal changes in patients with spondyloarthritis: comparison of MR imaging and radiographic appearances. Radiographics 2005;25:559–69. [6] Gaspersic N, Sersa I, Jevtic V, Tomsic M, Praprotnik S. Monitoring ankylosing spondylitis therapy by dynamic contrast-enhanced and diffusion-weighted magnetic resonance imaging. Skeletal Radiol 2008;37:123–31. [7] Hillengass J, Stieltjes B, Bäuerle T, et al. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and diffusion-weighted imaging of bone marrow in healthy individuals. Acta Radiol 2011;52:324–30. [8] Bozgeyik Z, Ozgocmen S, Kocakoc E. Role of diffusion-weighted MRI in the detection of early active sacroiliitis. Am J Roentgenol 2008;191:980–6. [9] Khoo MMY, Tyler PA, Saifuddin A, Padhani AR. Diffusion-weighted imaging (DWI) in musculoskeletal MRI: a critical review. Skeletal Radiol 2011;40:665–81. [10] Meurin A, Cernicanu A, Molinier S, et al. Diffusion-weighted MR imaging of the spine and cord. J Radiol 2010;91:352–66, quiz 367–68. [11] Dougados M, Gueguen A, Nakache JP, Nguyen M, Amor B. Evaluation of a functional index for patients with ankylosing spondylitis. J Rheumatol 1990;17:1254–5. [12] Dougados M, d’ Agostino M-A, Benessiano J, et al. The DESIR cohort: a 10year follow-up of early inflammatory back pain in France: study design and baseline characteristics of the 708 recruited patients. Joint Bone Spine 2011;78:598–603. [13] Kealey SM, Aho T, Delong D, Barboriak DP, Provenzale JM, Eastwood JD. Assessment of apparent diffusion coefficient in normal and degenerated intervertebral lumbar disks: initial experience. Radiology 2005;235:569–74. [14] Eguchi Y, Ohtori S, Yamashita M, et al. Diffusion magnetic resonance imaging to differentiate degenerative from infectious endplate abnormalities in the lumbar spine. Spine 2011;36:E198–202. [15] Horger M, Weisel K, Horger W, Mroue A, Fenchel M, Lichy M. Whole-body diffusion-weighted MRI with apparent diffusion coefficient mapping for early response monitoring in multiple myeloma: preliminary results. AJR Am J Roentgenol 2011;196:W790–5. [16] Hillengass J, Bäuerle T, Bartl R, et al. Diffusion-weighted imaging for noninvasive and quantitative monitoring of bone marrow infiltration in patients with monoclonal plasma cell disease: a comparative study with histology. Br J Haematol 2011;153:721–8. [17] Modic MT, Steinberg PM, Ross JS, Masaryk TJ, Carter JR. Degenerative disk disease: assessment of changes in vertebral body marrow with MR imaging. Radiology 1988;166:193–9. [18] Davis Jr JC, Mease PJ. Insights into the pathology and treatment of spondyloarthritis: from the bench to the clinic. In: Seminars in arthritis and rheumatism. 2008. p. 83–100. [19] Sieper J, Rudwaleit M, Baraliakos X, et al. The Assessment of SpondyloArthritis international Society (ASAS) handbook: a guide to assess spondyloarthritis. Ann Rheum Dis 2009;68(Suppl (2)):ii1–44. [20] Nguyen C, Bendeddouche I, Sanchez K, et al. Assessment of ankylosing spondylitis criteria in patients with chronic low back pain and vertebral endplate Modic I signal changes. J Rheumatol 2010;37:2334–9. [21] Messiou C, Collins DJ, Morgan VA, Desouza NM. Optimising diffusion weighted MRI for imaging metastatic and myeloma bone disease and assessing reproducibility. Eur Radiol 2011;21:1713–8. [22] Griffith JF, Yeung DKW, Antonio GE, et al. Vertebral marrow fat content and diffusion and perfusion indexes in women with varying bone density: MR evaluation. Radiology 2006;241:831–8.
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Comparison of apparent diffusion coefficient in spondylarthritis axial active inflammatory lesions and type 1 modic changes Benjamin Dallaudière a,b,c,∗ , Raphaël Dautry a,c,1 , Pierre-Marie Preux d,2 , Anne Perozziello c,e,3 , Julien Lincot a,1 , Elisabeth Schouman-Claeys a,c,1 , Jean-Michel Serfaty a,b,c,1 a
Hôpital Bichat – Claude Bernard, Service de Radiologie, 46, rue Henri Huchard, Paris 75018, France Hôpital Bichat – Claude Bernard, Inserm U698, 46, rue Henri Huchard, Paris 75018, France c Université Paris Diderot, Paris, France d Faculté de Médecine de Limoges, Unité fonctionelle de recherche clinique et de biostatistique, hôpital Le Cluzeau, 23, avenue Dominique Larrey, 87042 Limges Cedex, France e Hôpital Bichat – Claude Bernard, Unité de recherche clinique, 46, rue Henri Huchard, Paris 75018, France b
a r t i c l e
i n f o
Article history: Received 31 January 2013 Received in revised form 26 August 2013 Accepted 12 October 2013 Keywords: Spondylarthritis Spondylitis Diffusion weighted imaging Apparent diffusion coefficient Modic change
a b s t r a c t Objective: The goal of this study was to evaluate whether the values of ADC in spondylarthritis axial active inflammatory lesions are different from ADC values in type 1 Modic changes. Subjects and methods: 95 patients with recent lumbar pain, including 46 patients with diagnosed or suspected spondylarthritis and 49 patients with purely degenerative history, underwent spine MRI. T1w, STIR, and diffusion-weighted images (DWI) were obtained. Two musculoskeletal radiologists interpreted the images. Axial active inflammatory lesions from the SpA group and type 1 Modic changes from the degenerative group were identified on T1w and STIR sequences. ADC values from these lesions and from healthy subchondral bone were compared. Results: All axial active inflammatory lesions (n = 27) and type 1 Modic changes (n = 22) identified in T1w and STIR images were visible on DWI. ADC values were significantly higher (p < 0.05) for axial active inflammatory lesions (median = 0.788 × 10−3 mm2 /s, IQR 25–75 [0.7 × 10−3 mm2 /s; 0.9 × 10−3 mm2 /s]) than for type 1 Modic changes (median = 0.585 × 10−3 mm2 /s, IQR 25–75 [0.55 × 10−3 mm2 /s; 0.60 × 10−3 mm2 /s]) and normal subchondral bone (median = 0.443 × 10−3 mm2 /s, IQR 25–75 [0.40 × 10−3 mm2 /s; 0.50 × 10−3 mm2 /s]). Intra-class correlation coefficients for intra- and inter-reader ADC values comparison were excellent (0.89 and 0.98 respectively). Conclusion: DWI is a sensitive and fast sequence that offer the possibility of quantifying diffusion coefficients of the lesions, which could help to discriminate between spondylarthritis axial active inflammatory and type 1 Modic changes. © 2013 Elsevier Ireland Ltd. All rights reserved.
1. Introduction
Abbreviations: ADC, apparent diffusion coefficient; DWI, diffusion weighted imaging; ROI, region of interest; SpA, spondylarthritis; T1w, T1 weighted; ASAS/OMERACT, Assessment of Spondylarthritis International Society/Outcome Measures in Rheumatoid Arthritis Clinical Trials; DCE-MRI, dynamic contrastenhanced MRI. ∗ Corresponding author at: Hôpital Bichat – Claude Bernard, Service de Radiologie, 46, rue Henri Huchard, Paris 75018, France. Tel.: +33 1 40 25 80 80; fax: +33 1 40 25 88 73. E-mail addresses: [email protected], [email protected] (B. Dallaudière), raphael [email protected] (R. Dautry), [email protected] (P.-M. Preux), [email protected] (A. Perozziello), [email protected] (J. Lincot), [email protected] (E. Schouman-Claeys), [email protected] (J.-M. Serfaty). 1 Tel.: +33 1 40 25 80 80; fax: +33 1 40 25 88 73. 2 Tel.: +33 5 55 05 66 29. 3 Tel.: + 33 1 40 25 79 31. 0720-048X/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ejrad.2013.10.009
Lumbar pain is one of the most frequently reported symptoms, with a high impact on life quality [1]. When there is no evident traumatic, infectious or neoplastic context, back pain usually is of rheumatismal or degenerative cause. Most times clinical information in addition to imaging studies are sufficient to differentiate these two patterns. However the differential diagnosis can be tricky when clinical dates are scarce, and when lesions do not display a typical imaging aspect. Inflammatory spondylarthritis (SpA) are common diseases, whose frequency is estimated to be between 0.3 and 1.9% in the general population [1,2]. Because of the young age of the SpA patients, acute lumbar pains are usually attributed to SpA axial active inflammatory lesions but degenerative lesions such as type 1 Modic changes are a frequent cause for back pain and may be overlooked
B. Dallaudière et al. / European Journal of Radiology 83 (2014) 366–370
in young patients. Such misdiagnosis may lead to overtreatment of patients with either 3 months corticotherapy or immunosuppressive therapy. MRI, with T1 weighted and STIR sequences, is widely used to assess back pain patients. Differential diagnosis between axial active inflammatory lesions and Modic1 lesions can be tricky as lesion signal does not permit to discriminate between these two types of acute lesions [3,4], and anatomical topography sometimes overlap [5]. Recent works have evaluated new MRI technics for assessement and follow-up of spondylarthritis inflammatory lesion. Dynamic contrast-enhanced MRI (DCE-MRI) has been reported to be valuable in detecting early- and late-stage inflammation in the sacroiliac joints of patients with spondylarthropathy, as well as for the therapeutic follow-up of SpA [6,7]. Although in this specific clinical setting this technique has shown promising results, it is not routinely used in imaging of SpA. Another recent study has shown that diffusion weighted MRI (DWI) might be effective to differentiate acute degenerative lesions from early inflammatory sacro-iliitis [8]. Without the need of contrast injection, this makes DWI a potentially effective tool to differentiate axial active inflammatory lesions and type 1 Modic changes. To our knowledge, the role of diffusion weighted MRI with ADC in this application has not been studied [9,10]. The goal of our study was therefore to compare ADC values of axial active inflammatory lesions in spondylarthritis patients to ADC values of type 1 Modic changes in patients with a purely degenerative disease.
2. Materials and methods 2.1. Patients We conducted a mono-centric observational cohort study on 46 consecutive spondylarthritis patients and 49 purely degenerative patients from November 2010 to June 2012. For the spondylarthritis group, inclusion criteria were probable SpA (Amor score = 5) or confirmed SpA (Amor score ≥ 6) [11] presenting with a recent back pain (duration of less than a month). For the degenerative group, inclusion criteria were recent back pain with Amor score < 5 and no antecedent of personal of familial rheumatismal inflammatory disease. All patients were referred to our imaging department by the Rheumatology and Internal Medicine departments for an MRI of the spine. Exclusion criteria included contraindication to MRI (pregnancy, metallic implants, and claustrophobia) and suspicion of spine infection. Fifteen patients of the spondylarthritis group belonged to the DESIR study [12]. All patients were informed of the study procedure and gave their informed consent.
367
0 and 1000 s/mm2 (we chose this high b-value to enhance DWI sensitivity [8]). On the thoraco-lumbar spine, DWI parameters were: sagittal slices, TR 6000 ms, TE 1114 ms number of excitations, 2; duration 2.06 min, slice thickness, 5 mm; intersection gap, none. On the cervico-thoracic spine, DWI parameters were: axial slices, TR 7075 ms, TE 87.4 ms number of excitations, 2; duration 4.57 min, slice thickness, 5 mm; intersection gap, none. Axial slices were used to benefit from a low acquisition time and a lesser sensitivity to motion artifacts, as electrocardiogram (EKG) gated sagittal DWI are longer to acquire.
2.3. Image interpretation Patients’ identities were removed from all images. Two musculoskeletal radiologists blindly assessed images, in random patient order, for the presence of axial active inflammatory lesions and type 1 Modic changes. T1 and STIR weighted images were looked upon first, followed immediately by DWI analysis with ADC measurements on lesions identified on T1 and STIR images. Both readers, according to the same protocol, performed a second interpretation two weeks later. The presence of bone marrow edema (T1 hypo signal and STIR hyper signal) was considered as a marker of acute disease. ADC maps were obtained from DWI data, on a GE Healthcare Advantage Windows, 4.2 Workstation. In order to be more reproducible, measurements were made on the Kodac PACS system® using circular ROI having an area of at least 4 pixels, or 4 mm × 4 mm (information given by the PACS system) that was positioned on active inflammatory lesions and type 1 Modic changes, as well as on 2 different locations on normal subchondral bone.
2.4. Statistical methods Statistical analysis was performed using the SAS® software. ADC values for active inflammatory lesions, type 1 Modic changes and healthy subchondral bone were expressed as follows (median and inter quantile range (IQR)) and compared using a Wilcoxon signed rank sum test, for paired values, not normally distributed. ADC cut-off was determined using the boxplot graphic. We considered p < 0.05 as significant. Intra- and inter-reader intraclass correlation coefficients for ADC value measurement were calculated.
3. Results 3.1. Population
2.2. MRI All patients underwent MRI of the either the entire spine (n = 32 in the SpA group and n = 14 in the degenerative group) or the thoraco-lumbar spine (n = 14 in the SpA group and n = 35 in the degenerative group). MRI was performed using a 1.5-Tesla MR scanner (Twin Speed HDX: GE® Healthcare) with a spine coil (CTI Array by USAI; 6 elements, 6 channels). The following sequences were obtained: sagittal fast spin-echo T1-weighted (TR/TE: 660/9.5 ms, duration 2.53 min, FOV 48 cm, 13 slices), and sagittal STIR (TR/TE: 4700/68 ms, duration 3.36 min, FOV 48 cm, 13 slices) sequences. Slice thickness was 3 mm; number of excitations, 2; and intersection gap, none. In addition to these previous sequences, DWI (without suppression of fat signal) was performed using a single-shot spin echo-planar imaging sequence with diffusion gradient (b value) of
Ninety-five patients were included in this study, 46 in the spondylarthritis group (27 women) and 49 in the degenerative group (28 women). In the spondylarthritis group, mean age was 43 years (SD 5.3 years). Twenty-seven patients were clinically suspected of SpA and 19 patients were clinically diagnosed with SpA (including 14 patients with ankylosing spondylitis, 3 with psoriatic arthritis and 2 with spondylarthritis associated to inflammatory bowel disease). Mean back pain duration before MRI was 19 days (±3.2 days). Among the 27 patients with suspected SpA and the 19 patients with SpA, 0 and 15 respectively were receiving medical treatment (either anti-TNF or corticotherapy) at the time of imaging. In the degenerative group, mean age was 52 years (SD 4.1 years), mean back pain duration before MRI was 21 days (±2.8 days), and none had had therapeutic spinal injection.
368
B. Dallaudière et al. / European Journal of Radiology 83 (2014) 366–370
Fig. 1. (a)–(c) T1 (a), STIR (b), DWI (c) and ADC map (d) sagittal slices showing a rheumatismal acute L5-S1 spondylodiscitis.
Fig. 2. (a)–(c) T1 (a), STIR (b), DWI (c) and ADC map (d) sagittal slices showing a L4–L5 Modic 1 lesion.
3.2. Lesion detection by T1 and STIR weighted images In the spondylarthritis group, we found 27 typical axial active inflammatory lesions (17 anterior spondylitis, 2 discitis, 5 posterior arthritis and 3 costo-vertebral or costo-transversary arthritis) in 22 patients. In the degenerative group, we found 22 typical type 1 Modic changes in 14 patients, all located on vertebral endplates (20 on the lumbar spine and 2 on the cervical spine). 3.3. DWI study and ADC measurement All lesions identified on T1w and STIR was visible on DWI. ADC measurements were performed on 17 axial active inflammatory lesions (17 spondylitis, Fig. 1) out of 27 lesions detected, 22 type
1 Modic lesions (Fig. 2), and 92 subchondral bone with no lesion (Fig. 3). We excluded the 10 remaining axial active inflammatory lesions (posterior arthritis, costo-vertebral or costo-transversary arthritis lesions) from ADC analysis due to small size. ADC values were higher for axial active inflammatory lesions (median = 0.788 × 10−3 mm2 /s, IQR 25–75 [0.7 × 10−3 mm2 /s; 0.9 × 10−3 mm2 /s]) than for type 1 Modic changes (median = 0.585 × 10−3 mm2 /s, IQR 25–75 [0.55 × 10−3 mm2 /s; 0.60 × 10−3 mm2 /s]) and for normal subchondral bone (median = 0.445 × 10−3 mm2 /s, IQR 25–75 [0.40 × 10−3 mm2 /s; 0.50 × 10−3 mm2 /s]) (Fig. 4). Median ADC values were significantly different between axial active inflammatory lesions and Md1 (p = 0.0005), between axial active inflammatory lesions and normal subchondral bone (p < 0.0001) and between type 1 Modic changes and normal
Fig. 3. (a)–(c) T1 (a), STIR (a), DWI (b) and ADC map (c) sagittal slices of a normal lumbar spine.
B. Dallaudière et al. / European Journal of Radiology 83 (2014) 366–370
369
Fig. 4. Graphical view of the included patients, the identified lesions and the calculated ADC values.
subchondral bone (p = 0.0037). We used the boxplot graphic to visually determine the best ADC cut-off value to differentiate between axial active inflammatory lesions and type 1 Modic changes in our patients. The best ADC cut-off value was 0.58 × 10−3 mm2 /s (Fig. 5). ADC values were superior to 0.58 × 10−3 mm2 /s in 15 out of 17 inflammatory lesions in which ADC could be measured, and inferior to 0.58 × 10−3 mm2 /s in 19 out of 22 type 1 Modic changes. 3.4. Intra- and inter-reader agreement Inter-reader intraclass correlation coefficients for ADC measurements were very good (ICC = 0.98 and 0.89 respectively). Intra-reader agreement was very good as well, for both radiologists (ICC = 0.98 and 0.92 respectively). 4. Discussion Our results show for the first time that in case of recent back pain, ADC values of axial active inflammatory lesions in spondylarthritis patients are different from ADC values of type 1 Modic changes in patients with purely degenerative lesions. DWI is a fast MRI sequence, which offers without any contrast media injection, an approximation of micro vascular perfusion and diffusion of water molecules in the interstitium. Its role for assessment of various types of spine vertebral and discal lesions has been increasingly studied, in the field of degenerative [13],
Fig. 5. Graphical view of IQR 25–75 for ADC values for axial active inflammatory lesion, type 1 Modic changes and normal subchondral bone.
infectious [14] and neoplastic lesions [15,16]. Recently, Gaspersic et al. [6] have shown a correlation between clinical efficacy of ankylosing spondylitis treatment and decrease of ADC in corresponding inflammatory lesions, adding weight to the assumption that ADC value correlate to clinical intensity of inflammatory process. Bozgeyik et al. [8] have also observed that, in SpA patients, ADC values of sacro-iliitis lesions and subchondral sacro-iliac bone were different. Our results are consistent with these findings and suggest that ADC might be different between type 1 Modic changes and spondylarthritis inflammatory spine lesions. The pathophysiological substratum for the difference of ADC between SpA lesions and Type 1 Modic changes is unclear. In their original study, Modic et al. [17] analyzed histopathological sections of type 1 changes and found that they were associated with disruption and fissuring of endplates and formation of a fibrovascular granulation tissue. They concluded that type 1 changes correspond to the inflammatory stage of disk degenerative disease and indicate an active degenerative process. Concerning SpA lesions, although the underlying causes of inflammation are not well understood, a large body of evidence indicates that the pathology is at least partly mediated by immune cells and that TNF-␣ plays a central role in a majority of pathologic mechanisms that drive inflammation and damage. Several features of inflammatory arthritis, including angiogenesis and bone erosions without fissures, have been attributed to the extracellular matrix degradation by metallo-proteinases with also an increase in the number of osteoclasts present in the SpA bone [18]. The difference of ADC between these two types of lesions could be linked to the underlying mechanisms of inflammation, or could reflect a more severe inflammation in SpA lesions. The difference of ADC between SpA active inflammatory spine changes and type 1 Modic changes could be interesting in patients presenting non-specific bone lesions on MR images. Indeed, this differential diagnosis is a classical pitfall when interpreting inflammatory changes in spondylarthritis patient, and is sometime not feasible on conventional MRI results alone [19,20]. Our study has several limitations. First, in our population, the diagnostic of spondylarthritis was based on the Amor criteria, in order to have two clearly defined populations. However, the Amor criteria are not always fulfilled at the early stages of spondylarthritis, which is why it would be highly beneficial to be able to differentiate spondylarthritis lesions from Modic 1 changes using only MRI sequences. Second, we did not compare DWI/ADC to DCE-MRI. DCE-MRI has shown to better quantify inflammation than STIR sequences in the therapeutic follow-up of SpA [6], and so could be a good technique
370
B. Dallaudière et al. / European Journal of Radiology 83 (2014) 366–370
to differentiate spondylarthritis spine lesions and type 1 Modic changes. Neither did we compare DWI/ADC to the ASAS/OMERACT criteria. Indeed, the main goal of our study was to assess whether ADC values were different in the inflammatory lesions of spondylarthritis and in the type 1 Modic changes of our population. Based on these results as described above, our next step will be therefore to compare the diagnostic accuracy of DWI/ADC to other diagnostic tools such as the ASAS/OMERACT criteria and DCE-MRI. Third, a technical limitation is the low spatial resolution of the DWI sequence (in our study 1 mm × 1 mm with a 5 mm slice thickness) compared to the spatial resolution of T1w and STIR sequences. This explains our decision to exclude lesions smaller than 4 mm for ADC measurements. A solution to this problem could be to use DWI specifically on the lesion to characterize, using thinner slice, smaller FOV and systematic electrocardiogram (EKG) gated sequences. However, despite this limitation, we were able to find significant differences between our groups. Moreover, we found normal subchondral bone values consistent with the literature [21,22]. Future improvement in spatial resolution should help confirm differences in ADC measurement. Regarding the choice of b, we used a high diffusion gradient (b value) of 1000 s/mm2 to enhance the sensitivity of DWI [10] compared to lower b values [8]. Another limitation is related to the use of different axial slice orientation in cervico-thoracic acquisitions instead of the classical sagittal acquisition plan. This axial slice orientation was chosen as sagittal EKG-gated sequences (to avoid cardiac motion artifact) have a long acquisition time. Axial non EKG-gated acquisition almost avoids motion artifacts with a limited cost regarding time acquisition. Finally, the number of patients included in our study was limited and future prospective studies with patient and lesion follow-up will be needed to confirm our initial results. In conclusion, DWI is a sensitive, fast sequence and does not require a contrast agent. Our study suggests that DWI offer the possibility of quantifying diffusion coefficients of the lesions, which helps to discriminate between normal and involved active inflammatory and degenerative subchondral bone. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ejrad.2013. 10.009. References [1] Sieper J, Rudwaleit M, Khan MA, Braun J. Concepts and epidemiology of spondyloarthritis. Best Pract Res Clin Rheumatol 2006;20:401–17.
[2] Saraux A, Guillemin F, Guggenbuhl P, et al. Prevalence of spondyloarthropathies in France: 2001. Ann Rheum Dis 2005;64:1431–5. [3] Althoff CE, Feist E, Burova E, et al. Magnetic resonance imaging of active sacroiliitis: do we really need gadolinium. Eur J Radiol 2009;71:232–6. [4] Jevtic V, Kos-Golja M, Rozman B, McCall I. Marginal erosive discovertebral ‘Romanus’ lesions in ankylosing spondylitis demonstrated by contrast enhanced Gd-DTPA magnetic resonance imaging. Skeletal Radiol 2000;29:27–33. [5] Hermann K-GA, Althoff CE, Schneider U, et al. Spinal changes in patients with spondyloarthritis: comparison of MR imaging and radiographic appearances. Radiographics 2005;25:559–69. [6] Gaspersic N, Sersa I, Jevtic V, Tomsic M, Praprotnik S. Monitoring ankylosing spondylitis therapy by dynamic contrast-enhanced and diffusion-weighted magnetic resonance imaging. Skeletal Radiol 2008;37:123–31. [7] Hillengass J, Stieltjes B, Bäuerle T, et al. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and diffusion-weighted imaging of bone marrow in healthy individuals. Acta Radiol 2011;52:324–30. [8] Bozgeyik Z, Ozgocmen S, Kocakoc E. Role of diffusion-weighted MRI in the detection of early active sacroiliitis. Am J Roentgenol 2008;191:980–6. [9] Khoo MMY, Tyler PA, Saifuddin A, Padhani AR. Diffusion-weighted imaging (DWI) in musculoskeletal MRI: a critical review. Skeletal Radiol 2011;40:665–81. [10] Meurin A, Cernicanu A, Molinier S, et al. Diffusion-weighted MR imaging of the spine and cord. J Radiol 2010;91:352–66, quiz 367–68. [11] Dougados M, Gueguen A, Nakache JP, Nguyen M, Amor B. Evaluation of a functional index for patients with ankylosing spondylitis. J Rheumatol 1990;17:1254–5. [12] Dougados M, d’ Agostino M-A, Benessiano J, et al. The DESIR cohort: a 10year follow-up of early inflammatory back pain in France: study design and baseline characteristics of the 708 recruited patients. Joint Bone Spine 2011;78:598–603. [13] Kealey SM, Aho T, Delong D, Barboriak DP, Provenzale JM, Eastwood JD. Assessment of apparent diffusion coefficient in normal and degenerated intervertebral lumbar disks: initial experience. Radiology 2005;235:569–74. [14] Eguchi Y, Ohtori S, Yamashita M, et al. Diffusion magnetic resonance imaging to differentiate degenerative from infectious endplate abnormalities in the lumbar spine. Spine 2011;36:E198–202. [15] Horger M, Weisel K, Horger W, Mroue A, Fenchel M, Lichy M. Whole-body diffusion-weighted MRI with apparent diffusion coefficient mapping for early response monitoring in multiple myeloma: preliminary results. AJR Am J Roentgenol 2011;196:W790–5. [16] Hillengass J, Bäuerle T, Bartl R, et al. Diffusion-weighted imaging for noninvasive and quantitative monitoring of bone marrow infiltration in patients with monoclonal plasma cell disease: a comparative study with histology. Br J Haematol 2011;153:721–8. [17] Modic MT, Steinberg PM, Ross JS, Masaryk TJ, Carter JR. Degenerative disk disease: assessment of changes in vertebral body marrow with MR imaging. Radiology 1988;166:193–9. [18] Davis Jr JC, Mease PJ. Insights into the pathology and treatment of spondyloarthritis: from the bench to the clinic. In: Seminars in arthritis and rheumatism. 2008. p. 83–100. [19] Sieper J, Rudwaleit M, Baraliakos X, et al. The Assessment of SpondyloArthritis international Society (ASAS) handbook: a guide to assess spondyloarthritis. Ann Rheum Dis 2009;68(Suppl (2)):ii1–44. [20] Nguyen C, Bendeddouche I, Sanchez K, et al. Assessment of ankylosing spondylitis criteria in patients with chronic low back pain and vertebral endplate Modic I signal changes. J Rheumatol 2010;37:2334–9. [21] Messiou C, Collins DJ, Morgan VA, Desouza NM. Optimising diffusion weighted MRI for imaging metastatic and myeloma bone disease and assessing reproducibility. Eur Radiol 2011;21:1713–8. [22] Griffith JF, Yeung DKW, Antonio GE, et al. Vertebral marrow fat content and diffusion and perfusion indexes in women with varying bone density: MR evaluation. Radiology 2006;241:831–8.
