Session Information
Date: Sunday, October 21, 2018
Session Type: ACR Poster Session A
Session Time: 9:00AM-11:00AM
The ASAS MRI group has generated updated consensus lesion definitions (ASAS_MRI_defn) for the spectrum of MRI lesions in the SIJ. Their relative performance as diagnostic indicators requires assessment to understand their value in clinical practice when assessing patients referred with undiagnosed back pain. We aimed to determine optimal quantitative SPARCC cut-offs for specific MRI lesions reflecting diagnosis of axSpA in patients with undiagnosed back pain recruited to the ASAS Classification Cohort (ASAS-CC)1 and diagnosed by local rheumatologist expert opinion.
ASAS_MRI_defn were recorded in an eCRF that comprises global assessment (active or structural lesion typical of axSpA present/absent), links to reference images, and detailed scoring of lesions per SIJ quadrant (SPARCC SIJ inflammation, SPARCC SIJ structural). MRI images were available from 278 of the 495 cases that had MRI performed in the ASAS-CC and were evaluated by 7 experienced readers from the ASAS-MRI group. Detailed SPARCC scoring data was based only on assessment of images in DICOM format. We calculated sensitivity and specificity for varying numbers of SIJ quadrants with bone marrow edema (BME), erosion, and fatty lesions for diagnosis of axSpA as determined by the local rheumatologist at baseline and after a mean 4.4 years of follow up. Lesion cut-offs were tested according to data based on majority agreement (≥4/7 readers) and from any 2 central readers.
Lesion cut-offs based on BME were most sensitive for diagnosis at baseline and follow up but erosion- and fatty-lesion based cut-offs were more specific (Table). There was a consistent improvement in the sensitivity and specificity performance of cut-offs based on erosions and fatty lesions according to diagnosis at baseline and then at follow up after 4.4 years but not for cut-offs based on BME. This improvement was also consistently observed in data based on majority agreement of central readers. Sclerosis based cut-offs performed least well but specificity was still 90% at follow up.
The ASAS consensus definitions for active and structural lesions and scoring cut-offs based on the presence of these lesions in 2-3 SIJ quadrants have comparable diagnostic performance. However, diagnostic performance consistently improves for erosion and fatty lesion based cut-offs after follow up.
1. Rudwaleit et al. Ann Rheum Dis 2009;68: 777-831.
Table. Sensitivities and specificities of cut-offs for SIJ lesion scores (number of SIJ quadrants) according to presence of axSpA according to diagnostic ascertainment of local physician in ASAS classification study.
|
MRI lesion cut-offs* |
axSpA at Baseline (n 237) |
axSpA at Follow up (n 143) |
||
|
Sensitivity (95% CI) |
Specificity (95% CI) |
Sensitivity (95% CI) |
Specificity (95% CI) |
|
|
Any 2 central reader data |
||||
|
BME Score ≥2 |
50.00 (40.6 – 59.4) |
82.46 (70.1 – 91.3) |
46.48 (34.5 – 58.7) |
86.36 (65.1 – 97.1) |
|
BME Score ≥3 |
40.52 (31.5 – 50.0) |
94.74 (85.4 – 98.9) |
33.80 (23.0 – 46.0) |
90.91 (70.8 – 98.9) |
|
Erosion Score ≥2 |
35.34 (26.7 – 44.8) |
92.98 (83.0 – 98.1) |
33.80 (23.0 – 46.0) |
95.45 (77.2 – 99.9) |
|
Erosion Score ≥3 |
31.03 (22.8 – 40.3) |
92.98 (83.0 – 98.1) |
33.80 (23.0 – 46.0) |
95.45 (77.2 – 99.9) |
|
Fatty lesion (any) ≥2 |
34.48 (25.9 – 43.9) |
92.98 (83.0 – 98.1) |
38.03 (26.8 – 50.3) |
95.45 (77.2 – 99.9) |
|
Fatty lesion (any) ≥3 |
31.90 (23.6 – 41.2) |
94.74 (85.4 – 98.9) |
35.21 (24.2 – 47.5) |
95.45 (77.2 – 99.9) |
|
Sclerosis ≥2 |
32.76 (24.3 – 42.1) |
82.46 (70.1 – 91.3) |
29.58 (19.3 – 41.6) |
90.91 (70.8 – 98.9) |
|
Sclerosis ≥3 |
28.45 (20.5 – 37.6) |
82.46 (70.1 – 91.3) |
26.76 (16.9 – 38.6) |
90.91 (70.8 – 98.9) |
|
Majority agreement ((≥4/7) central reader data |
||||
|
BME Score ≥2 |
36.21 (27.5 – 45.6) |
92.98 (83.0 – 98.1) |
29.58 (19.3 – 41.6) |
90.91 (70.8 – 98.9) |
|
BME Score ≥3 |
32.76 (24.3 – 42.1) |
100.00 (93.7 – 100.0) |
28.17 (18.1 – 40.1) |
95.45 (77.2 – 99.9) |
|
Erosion Score ≥2 |
26.72 (18.9 – 35.7) |
96.49 (87.9 – 99.6) |
29.58 (19.3 – 41.6) |
100.00 (84.6 – 100.0) |
|
Erosion Score ≥3 |
24.14 (16.7 – 33.0) |
96.49 (87.9 – 99.6) |
29.58 (19.3 – 41.6) |
100.00 (84.6 – 100.0) |
|
Fatty lesion (any) ≥2 |
20.69 (13.7 – 29.2) |
98.25 (90.6 – 100.0) |
21.13 (12.3 – 32.4) |
100.00 (84.6 – 100.0) |
|
Fatty lesion (any) ≥3 |
18.97 (12.3 – 27.3) |
98.25 (90.6 – 100.0) |
21.13 (12.3 – 32.4) |
100.00 (84.6 – 100.0) |
|
Sclerosis ≥2 |
18.10 (11.6 – 26.3) |
91.23 (80.7 – 97.1) |
18.31 (10.1 – 29.3) |
90.91 (70.8 – 98.9) |
|
Sclerosis ≥3 |
16.38 (10.2 – 24.4) |
91.23 (80.7 – 97.1) |
15.49 (8.0 – 26.0) |
90.91 (70.8 – 98.9) |
* Based on SPARCC scoring and number of SIJ quadrants
To cite this abstract in AMA style:
Maksymowych WP, Baraliakos X, Lambert RG, Machado P, Sieper J, Wichuk S, Poddubnyy D, Pedersen SJ, Paschke J, Østergaard M, Weber U. Diagnostic Performance of MRI Lesions in the Sacroiliac Joints According to Updated Assessments in Spondyloarthritis International Society Lesion Definitions: A Central Reader Assessment of MRI Scans from the Assessments in Spondyloarthritis Classification Cohort [abstract]. Arthritis Rheumatol. 2018; 70 (suppl 9). https://acrabstracts.org/abstract/diagnostic-performance-of-mri-lesions-in-the-sacroiliac-joints-according-to-updated-assessments-in-spondyloarthritis-international-society-lesion-definitions-a-central-reader-assessment-of-mri-scans/. Accessed .« Back to 2018 ACR/ARHP Annual Meeting
ACR Meeting Abstracts - https://acrabstracts.org/abstract/diagnostic-performance-of-mri-lesions-in-the-sacroiliac-joints-according-to-updated-assessments-in-spondyloarthritis-international-society-lesion-definitions-a-central-reader-assessment-of-mri-scans/