TOPLINE:

Combined definitions for the MRI of sacroiliac joint (MRI-SIJ) lesions proposed by Leiden and the Assessment of Spondyloarthritis International Society (ASAS)-MRI study group demonstrated a high specificity and positive predictive value (PPV) in early axial spondyloarthritis (axSpA). Structural lesions were infrequent in early disease, contributing only 6%-11% additional sensitivity beyond active lesion definitions alone.

METHODOLOGY:

• Researchers analysed baseline and 2-year follow-up data from the SPACE inception cohort, which included patients aged 16 years or older with chronic back pain of unknown origin lasting 3 months to 2 years from centres in Netherlands, Norway, Sweden, and Italy.
• A total of 643 patients (mean age, 30 years; SD, 8 years; 39% males) with a definite or most likely diagnosis of axSpA and non-axSpA after 2 years were included (335 with axSpA and 308 with non-axSpA), with a mean symptom duration of 13 months (SD, 7 months).
• Three central readers independently scored baseline MRI-SIJ for structural lesions, including erosions and fat lesions, and active lesions such as bone marrow oedema.
• Definitions for structural and active lesions proposed by Leiden (2016) and the ASAS-MRI study group (2021) were assessed for predictive validity against the 2-year rheumatologist diagnosis, with validation requiring a specificity of at least 95% and a PPV of at least 95%.
• Outcome measures included sensitivity, specificity, PPV, and negative predictive value for each MRI-SIJ lesion definition.

TAKEAWAY:

• All individual Leiden structural lesion definitions met the validation threshold; however, most MRI study group structural lesion definitions were validated, except for the definition of two or more consecutive slices affected by erosions (specificity, 98.4%; PPV, 91.2%).
• The ASAS-MRI-SIJ+ definition met the validation threshold and demonstrated a higher sensitivity than the MRI study group active lesion definition (39.7% vs 30.7%), with a similar specificity (98.1% vs 99.4%).
• The combination of the Leiden structural lesion definition with the ASAS-MRI-SIJ+ definition met the validation threshold with the highest sensitivity (45.7%), whereas the overall MRI study group structural lesion definition did not meet the validation threshold (specificity, 98.1%; PPV, 93.0%) in any reader.
• Structural lesions were infrequent in early axSpA (2%-14% of patients) and, when combined with active lesions, increased sensitivity by only 6%-11% beyond the 31%-40% sensitivity of active lesion definitions alone.

IN PRACTICE:

“All MRI-SIJ lesion definitions proposed by Leiden and most proposed definitions by the MRI study group were validated in early axSpA, demonstrating high specificity and PPV. Structural lesions were infrequent, highlighting the limited role of structural changes in early diagnosis,” the authors wrote.

“The ASAS-MRI-SIJ+ definition outperformed the MRI study group active lesion definition, and combining structural and active lesions provided only a small increase in sensitivity of early disease,” they added.

SOURCE:

This study was led by Liese J.E. de Bruin, Leiden University Medical Center, Leiden, Netherlands. It was published online on March 09, 2026, in RMD Open.

LIMITATIONS:

A substantial overlap (43%) of patients with the cohort from the Leiden study was observed, which could have affected the generalisability of the findings; however, over 300 additional patients and two new centres were included with similar results. The study did not investigate the anatomic distribution, frequency, or size of the MRI-SIJ lesions, which could have provided additional diagnostic value in early axSpA. The study did not investigate MRI-SIJ lesions in combination with other SpA features, which may have limited the understanding of their diagnostic value in clinical practice.

DISCLOSURES:

The authors did not declare any specific grant for this research from any funding agency in the public, commercial, or not-for-profit sectors. Several authors reported receiving travelling fees, grants, consultancy fees, or advisory board fees from several pharmaceutical companies including but not limited to UCB, MSD, Novartis, AbbVie, EULAR, Nordic Pharma, FWRO-KBS, and ASAS. One author reported being the director of Imaging Rheumatology bv.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.