Session Information
Date: Monday, November 9, 2015
Title: Imaging of Rheumatic Diseases Poster II: X-ray, MRI, PET and CT
Session Type: ACR Poster Session B
Session Time: 9:00AM-11:00AM
Background/Purpose: Axial spondyloarthritis (axSpA) can be difficult to diagnose at an early stage of the disease. Identification of structural lesions of the sacroiliac joints (SIJ) on X-rays is often delayed for several years and remains very disputable, even for trained readers. Inflammatory lesions on MRI are taken into account in classification criteria of axSpA but the validity of MRI in detecting structural lesions in clinical practice has not been assessed. The objective of this study was to evaluate the reliability of recognising structural lesions on MRI-SIJ in patients with inflammatory back pain (IBP) suggestive of axSpA in clinical practice compared to a central reading.
Methods: Patients aged 18-50, with recent (<3 years) and chronic (≥3 months) IBP, suggestive of axSpA were included in the DEvenir des Spondyloarthrites Indifférenciées Récentes (DESIR) cohort. MRI-SIJ structural lesions (erosions, fatty lesions, ankylosis) were scored by a local reader (non specifically trained radiologist or rheumatologist, with access to clinical data) and by two trained central readers (blinded for clinical data and other imaging modalities). Local readers scored each SIJ in 3 grades (0=normal, 1=doubtful, 2=definite) regarding the presence of structural lesions. A method similar to the SPARCC SIJ Structural Score was used for central reading.
The analyses were done based on the individual scores as well as the mean score of the two central readers.Agreement was calculated (kappa; % positive and negative agreement) between central readers regarding 9 possible definitions of a positive MRI-SIJ (table), and between local readers (3 definitions).
Results: In this analysis, 664/708 patients with complete available images were included. According to the central reading 39% had a least 1 erosion, fatty lesion or (partial) ankylosis, and according to the local reading, 35.4% patients had a positive MRI-SIJ based on structural lesions(at least unilateral ‘doubtful’ or ‘definite structural lesion’). Agreement between central readers was “moderate” for 4 definitions MRI-SIJ structural lesions, and “fair” for the other 5 definitions (data not shown).When comparing local to central reading (both definitions 1) as external standard (table, mean score central readers), there were 13.1% false positive (overclassified) and 16.7% false negative classifications (not detected).Agreements improve when adding fat lesions to erosions in the definition of a positive MRI-SIJ.
Conclusion: In this study, agreement on positive MRI-SIJ based on the presence of structural lesions was fair to moderate between local and central readers, as well as between trained central readers. The reliability improved when fat lesions are combined with erosions. The diagnostic value of MRI structural lesions needs to be determined in longitudinal follow-up.
|
|
Definition 1: ≥1 erosion, fatty lesion or (partial) ankyosis |
Definition 2: ≥1 erosion |
Definition 3: ≥2 erosions |
|||
|
MRI Local Reading |
PPA |
NPA |
PPA |
NPA |
PPA |
NPA |
|
Definition 1 : At least 1 SI-joint with doubtful or definite structural lesions |
59,9 |
76,3 |
57,8 |
77,2 |
48,6 |
80,5 |
|
K(95%CI): 0,36 (0,29 to 0,44) |
K(95%CI): 0,35 (0,27 to 0,43) |
K(95%CI): 0,31 (0,24 to 0,39) |
||||
|
Definition 2 : At least 1SI-joint with definite structural lesions |
57,6 |
79,5 |
54,3 |
79,9 |
49,8 |
85,0 |
|
K(95%CI): 0,38 (0,31 to 0,46) |
K(95%CI): 0,35 (0,27 to 0,43) |
K(95%CI): 0,35 (0,27 to 0,43) |
||||
|
Definition 3 : Bilateral definite structural lesions |
51,6 |
80,6 |
48,0 |
80,8 |
50,0 |
88,3 |
|
K(95%CI): 0,36 (0,29 to 0,42) |
K(95%CI): 0,31 (0,24 to 0,39) |
K(95%CI): 0,38 (0,30 to 0,47) |
||||
|
|
Definition 4: ≥3 erosions |
Definition 5: ≥1 fatty lesion |
Definition 6: ≥2 fatty lesions |
|||
|
Definition 1 |
38,0 |
80,2 |
55,9 |
79,0 |
52,3 |
81,8 |
|
K(95%CI): 0,24 (0,17 to 0,30) |
K(95%CI):0,35 (0,28 to 0,43) |
K(95%CI): 0,36 (0,29 to 0,43) |
||||
|
Definition 2 |
40,8 |
85,6 |
56,5 |
83,3 |
54,2 |
86,3 |
|
K(95%CI): 0,28 (0,20 to 0,37) |
K(95%CI):0,40 (0,32 to 0,48) |
K(95%CI): 0,41 (0,33 to 0,49) |
||||
|
Definition 3 |
46,4 |
90,1 |
52,1 |
85,1 |
56,9 |
89,9 |
|
K(95%CI): 0,37 (0,28 to 0,46) |
K(95%CI): 0,38 (0,30 to 0,46) |
K(95%CI):0,47 (0,38 to 0,55) |
||||
|
|
Definition 7: ≥3 fatty lesions |
Definition 8: (partial) ankylosis |
Definition 9: ≥3 erosions and/or fatty lesions |
|||
|
Definition 1 |
46,9 |
82,9 |
19,0 |
74,8 |
59,3 |
79,0 |
|
K(95%CI): 0,34 (0,27 to 0,41) |
K(95%CI): 0,01 (-0,05 to 0,08) |
K(95%CI):0,38 (0,31 to 0,46) |
||||
|
Definition 2 |
50,2 |
87,7 |
19,7 |
81,0 |
58,8 |
82,8 |
|
K(95%CI): 0,40 (0,31 to 0,48) |
K(95%CI): 0,04 (-0,03 to 0,11) |
K(95%CI):0,42 (0,35 to 0,49) |
||||
|
Definition 3 |
56,2 |
91,8 |
19,9 |
85,7 |
54,4 |
84,4 |
|
K(95%CI):0,48 (0,39 to 0,57) |
K(95%CI): 0,06 (-0,02 to 0,15) |
K(95%CI): 0,41 (0,33 to 0,48) |
||||
PPA/NPA= positive/negative percent agreement; K=Kappa
To cite this abstract in AMA style:
Jacquemin C, Rubio Vargas R, van den Berg R, Ferkal S, Le Corvoisier P, Rahmouni A, Loeuille D, van der Heijde D, Claudepierre P. What Is the Reliability of Recognizing Structural MRI Lesions of Sacroiliac Joints in Patients with Recent Inflammatory Back Pain in Clinical Practice? [abstract]. Arthritis Rheumatol. 2015; 67 (suppl 10). https://acrabstracts.org/abstract/what-is-the-reliability-of-recognizing-structural-mri-lesions-of-sacroiliac-joints-in-patients-with-recent-inflammatory-back-pain-in-clinical-practice/. Accessed .« Back to 2015 ACR/ARHP Annual Meeting
ACR Meeting Abstracts - https://acrabstracts.org/abstract/what-is-the-reliability-of-recognizing-structural-mri-lesions-of-sacroiliac-joints-in-patients-with-recent-inflammatory-back-pain-in-clinical-practice/