What Are the Optimal MRI Lesion Cut-Offs That Define Active and Structural Lesions in the Sacroiliac Joint As Being Typical of Axial Spondyloarthritis By Expert Readers?

Walter P. Maksymowych^1,2, Pedro Machado³, Ulrich Weber⁴, Xenofon Baraliakos⁵, Joachim Sieper⁶, Stephanie Wichuk¹, Denis Poddubnyy⁶, Mikkel Østergaard⁷, Joel Paschke², Robert G. Lambert¹ and Susanne J Pedersen⁷, ¹University of Alberta, Edmonton, AB, Canada, ²CaRE Arthritis, Edmonton, AB, Canada, ³University College London, London, United Kingdom, ⁴University of Southern Denmark, Odense, Denmark, ⁵Rheumazentrum Ruhrgebiet Herne, Herne, Germany, ⁶Charité Universitätsmeidzin Berlin, Berlin, Germany, ⁷COPECARE University of Copenhagen, Copenhagen, Denmark

Meeting: 2018 ACR/ARHP Annual Meeting

Keywords: classification criteria and spondylarthritis, MRI

Session Information

Date: Wednesday, October 24, 2018

Session Title: 6W011 ACR Abstract: Spondyloarthritis Incl PsA–Clinical VI: Imaging of Axial SpA (2922–2927)

Session Type: ACR Concurrent Abstract Session

Session Time: 9:00AM-10:30AM

Background/Purpose:

Defining what constitutes a definite MRI lesion in the SIJ typical of axSpA has been hampered by lack of international consensus for standardized lesion descriptions and definitions. The ASAS MRI group has generated updated consensus lesion definitions (ASAS_MRI_defⁿ) for the spectrum of MRI lesions in the SIJ and these have been applied to the assessment of MRI images from the ASAS Classification Cohort (ASAS-CC) by 7 experts in MRI interpretation from the ASAS-MRI group. We aimed to determine optimal quantitative SPARCC cut-offs for specific MRI lesions for a definite MRI lesion typical of axSpA, the external reference being majority reader decision as to the presence of an active or structural lesion typical of axSpA.

Methods:

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. Detailed SPARCC scoring data was based only on assessment of images in DICOM format (n =175). We calculated sensitivity and specificity for varying numbers of SIJ quadrants with bone marrow edema (BME), erosion, and fatty lesions where a majority of readers (≥4/7) agreed as to the presence of an active or structural lesion typical of axSpA.

Results:

BME was the most frequent lesion (SPARCC BME score, mean (SD) =14.4(15.5) in cases where a majority of readers agreed there was an active lesion typical of axSpA. For majority agreement as to the presence of a structural lesion typical of axSpA, the most frequent lesions were fatty lesion, and erosion (mean (SD) SPARCC score per case of 8.6(9.1) and 8.2(5.7), respectively). Other structural lesions were much less frequent. Optimal SPARCC cut-offs that focus on specificity (≥95%) were BME ≥3 for defining an active lesion, and fatty lesion ≥3 and erosion ≥2 for defining a structural lesion (Table).

Conclusion:

The ASAS MRI specific lesion definitions perform well when judged against the expert opinion of ASAS-MRI readers. In particular, the optimal cut-off for an erosion lesion, as defined by ASAS, has similar performance to the optimal cut-off for BME.

Table. Sensitivities and specificities of cut-offs for SIJ lesion scores (number of SIJ quadrants) according to presence of definite active or structural lesion typical for axSpA (≥4/7 readers in agreement)

	Sensitivity (95% CI)	Specificity (95% CI)
Majority agree active lesions indicative of axSpA present
BME Score ≥2	80.0 (65.4 – 90.4)	93.9 (88.2 – 97.3)
BME Score ≥3	77.8 (62.9 – 88.8)	99.2 (95.8 – 100.0)
Majority agree structural lesions indicative of axSpA present
Fatty lesion (any) ≥2	63.2 (46.0 – 78.2)	90.5 (83.7 – 95.2)
Fatty lesion (any) ≥3	55.3 (38.3 – 71.4)	94.8 (89.1 – 98.1)
Erosion Score ≥2	84.2 (68.7 – 94.0)	96.6 (91.4 – 99.1)
Erosion Score ≥3	79.0 (62.7 – 90.4)	99.1 (95.3 – 100.0)
Majority agree active OR structural lesions indicative of axSpA present
BME Score ≥2	70.4 (56.4 – 82.0)	95.0 (89.5 – 98.2)
BME Score ≥3	66.7 (52.5 – 78.9)	100.0 (97.0 – 100.0)
Fatty lesion ≥2	50.0 (36.1 – 63.9)	93.4 (87.4 – 97.1)
Fatty lesion ≥3	40.7 (27.6 – 55.0)	95.9 (90.6 – 98.6)
Erosion Score ≥2	63.0 (48.7 – 75.7)	98.4 (94.2 – 99.8
Erosion Score ≥3	57.4 (43.2 – 70.8)	100.0 (97.0 – 100)

Disclosure: W. P. Maksymowych, CaRE rthritis, 9; P. Machado, None; U. Weber, None; X. Baraliakos, None; J. Sieper, None; S. Wichuk, None; D. Poddubnyy, None; M. Østergaard, None; J. Paschke, None; R. G. Lambert, None; S. J. Pedersen, None.

To cite this abstract in AMA style:

Maksymowych WP, Machado P, Weber U, Baraliakos X, Sieper J, Wichuk S, Poddubnyy D, Østergaard M, Paschke J, Lambert RG, Pedersen SJ. What Are the Optimal MRI Lesion Cut-Offs That Define Active and Structural Lesions in the Sacroiliac Joint As Being Typical of Axial Spondyloarthritis By Expert Readers? [abstract]. Arthritis Rheumatol. 2018; 70 (suppl 10). https://acrabstracts.org/abstract/what-are-the-optimal-mri-lesion-cut-offs-that-define-active-and-structural-lesions-in-the-sacroiliac-joint-as-being-typical-of-axial-spondyloarthritis-by-expert-readers/. Accessed January 26, 2020.

« Back to 2018 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/what-are-the-optimal-mri-lesion-cut-offs-that-define-active-and-structural-lesions-in-the-sacroiliac-joint-as-being-typical-of-axial-spondyloarthritis-by-expert-readers/