Muscle Imaging in Inclusion Body Myositis: Refinement of MRI Criteria and Insights Into Upper Body Involvement

Background: The diagnosis of inclusion body myositis (IBM) can be delayed because of its heterogeneous clinical presentation and the lack of specific biomarkers. Muscle imaging has gained increasing relevance over the past decade and is now included among the supportive criteria in the international diagnostic guidelines. This study aimed to refine MRI criteria for IBM to facilitate clearer pattern recognition, increase their reproducibility and broader clinical applicability. We also aimed to provide a comprehensive evaluation of muscle wasting across the entire body, including less frequently assessed regions such as the neck, scapular girdle and trunk muscles, and to explore the presence of radiological IBM phenotypes through cluster analysis. Methods: Sixty-eight MRI scans and clinical records from patients diagnosed with IBM between 2003 and 2024 (60% males; mean age: 66 years, range: 46–85) were retrospectively reviewed. We defined a new set of three main and three supportive MRI criteria based on muscle imaging features and assessed their sensitivity. Whole body muscle involvement was semi-quantitatively scored using standardized scales across 6006 muscles. Pairwise correlation and K-means cluster analysis were performed to evaluate clinical–radiological relationships and identify phenotypic subgroups. Results: The revised MRI criteria achieved 96% sensitivity. Performance was consistent across clinical subgroups and remained robust (83%) in patients with atypical onset. Whole-body analysis highlighted mild but frequent wasting of paraspinal (90% of scans) and neck and scapular girdle muscles (87%), while intracranial muscles were consistently unaffected. Correlation analysis underlined a significant association between radiological and functional involvement in the lower (r = 0.57, p < 0.001) but not in the upper body. Cluster analysis revealed two overlapping but distinguishable imaging phenotypes, characterized by different involvement of paraspinal and distal leg muscles. Cluster 1 showed a higher proportion of male patients. Conclusions: The revised MRI criteria allow a reliable and easy recognition of the IBM pattern of muscle involvement, while whole-body imaging offers additional insights into disease heterogeneity and supports patient stratification in clinical trials. Clustering results also highlighted a possible sex-related influence on muscle vulnerability. The observed clinico-radiological correlations further support the role of muscle MRI indices as a surrogate outcome measure of muscle function.