Arthritis & Rheumatism, Volume 62,
November 2010 Abstract Supplement

Abstracts of the American College of
Rheumatology/Association of Rheumatology Health Professionals
Annual Scientific Meeting
Atlanta, Georgia November 6-11, 2010.

A Novel Tool for Fully Automatic Quantification of MRI Synovitis in Rheumatoid Arthritis Is More Sensitive to Change Than the Current Standard Semi-Quantitative Assessment.

Bowes1,  Michael A., Freeston4,  Jane E., Vital4,  Edward J., Vincent2,  Graham, Guillard2,  Gwenael, Bird5,  Paul, Emery3,  Paul

Imorphics Ltd, Manchester, United Kingdom
Imorphics Ltd
Section of Musculoskeletal Disease, University of Leeds, UK
University of NSW, Sydney, Australia


Current MRI evaluation in rheumatoid arthritis (RA) employs a semi-quantitative tool (OMERACT RAMRIS). Fully automated quantification of change in MRI pathology remains a highly desirable outcome for clinical trials and practice. 3D statistical appearance models have been used in osteoarthritis to detect very small physical changes in bone, cartilage and soft tissue.


To determine the responsiveness of a novel 3D automated measurement of MRI synovitis in the MCP and wrist joints of RA subjects in an open label study and to compare this with the RAMRIS synovitis assessment.


3T VIBE MR images pre- and post-contrast of 47 established RA patients who had recently commenced the same biological therapy were evaluated. Subjects were imaged at baseline and regular intervals after 6 months. 3D models or 'masks' of the joints assessed by RAMRIS were created from a separate dataset of RA images. Masks were created for the MCP joints, the radio-carpal joint, the radio-ulnar joint, and a mask for the carpal bones. These masks were fitted to each pre-contrast image and then subtraction images were created using a shuffle transform method. Synovitis volume (in mm3) was calculated automatically from the enhancing voxels within each mask. RAMRIS scoring was performed by a single experienced reader, blinded to time point.


Mean age of patients was 54.7 years, 88% women, 88.9% RF positive, mean number of previous DMARDs 2.88, 22/55 patients had received previous anti-TNF [10 1TNF, 9 2TNFs, 3 3TNFs] therapy. The numbers of patients with scans available at each time point were: 46 (6 months), 35, 15, 21, and 14 (14 months). At the group level, significant changes in automated synovitis volume were demonstrated at all time points except 10 months (where the size of the cohort was 15 individuals). RAMRIS scoring showed a similar pattern of change but with no significant change at any time point (Figure 1).

Figure 1. Change in synovitis volume (ml) and RAMRIS scores from baseline, with 95% confidence limits. Cohort size at each time point is shown on left hand graph. RAMRIS synovitis change scores did not achieve statistical significance at any time point. *= p<0.05, **= p<0.005.


Fully automated quantitative synovitis measurement is now feasible and practical, and in this observational study proved to be more responsive than semi-quantitative scores. This method could also be applied to clinical practice as well as clinical trials, as synovial volume is measured at the individual joint level.

To cite this abstract, please use the following information:
Bowes, Michael A., Freeston, Jane E., Vital, Edward J., Vincent, Graham, Guillard, Gwenael, Bird, Paul, et al; A Novel Tool for Fully Automatic Quantification of MRI Synovitis in Rheumatoid Arthritis Is More Sensitive to Change Than the Current Standard Semi-Quantitative Assessment. [abstract]. Arthritis Rheum 2010;62 Suppl 10 :118
DOI: 10.1002/art.27887

Abstract Supplement

Meeting Menu