Abstract Magnetic resonance imaging (MRI) is a cornerstone in the evaluation and monitoring of axial spondyloarthritis (axSpA), a chronic inflammatory condition primarily affecting the sacroiliac joints (SIJs), spine, entheses, and peripheral joints. Accurate quantification of axSpA-related changes in MRI is critical for effective research and patient management. However, current lesion detection and grading assessments suffer from substantial intra- and inter-rater variability, limiting their consistency and reliability. This study addresses these challenges by focusing on automated lesion detection in SIJ MRI to enhance accuracy and reduce variability. Our key contributions include: (i) developing a fully automated pipeline for detecting five distinct MRI lesion types (Bone Marrow Oedema, Ankylosis, Sclerosis, Erosions, Fatty Lesions) in the SIJ, (ii) validating the approach on a completely independent dataset, and (iii) proposing a simple approach to learn a classification model from multiple readings or labels for a given sample.
Jamaludin et al. (Tue,) studied this question.