Use of multispectral magnetic resonance imaging analysis to quantify erosive changes in the hands of patients with rheumatoid arthritis: Short-term and long-term longitudinal studies
✍ Scribed by Susanne Ostrowitzki; János Rédei; John A. Lynch; Richard A. D. Carano; Souhil Zaim; Yves Miaux; Harry K. Genant
- Publisher
- John Wiley and Sons
- Year
- 2004
- Tongue
- English
- Weight
- 348 KB
- Volume
- 50
- Category
- Article
- ISSN
- 0004-3591
No coin nor oath required. For personal study only.
✦ Synopsis
Abstract
Objective
Magnetic resonance imaging (MRI) has been shown to be more sensitive than radiography for detecting bone erosions in rheumatoid arthritis (RA). Semiquantitative scoring based on visual image assessment has been introduced. However, there is considerable interest in true quantitative measures, particularly in the context of clinical trials designed to show differences between treatment groups. This study was undertaken to investigate the use of a new quantitative approach, multispectral (MS) image analysis, for assessing erosive change.
Methods
T1‐weighted spin‐echo (SE) and fat‐suppressed gradient‐echo (GE) sequences of metacarpophalangeal joints of the dominant hand were acquired at various time points throughout a 2‐year period. MS analysis was applied to all images, resulting in segmentation into a generalized bone and a soft tissue class. Voxel changes from one to the other class identified apparent bone lesion volume change (ΔBLV). MR images were also visually scored for erosions (E score). All analyses were performed separately, on a per‐joint basis, for short‐term and long‐term data sets.
Results
Analysis of variance with adjustment for individual effect revealed similar results in the short‐term and the long‐term studies, using either GE or SE images for visual assessment. Patients with an increase in E score on visual assessment had a significantly higher ΔBLV than those without.
Conclusion
Temporal MS analysis of MRIs can be used to detect and quantify erosive changes in RA. This semiautomated method may be useful for demonstrating differences between treatment groups in clinical trials.