Simplified approach to MR image quantification of the rheumatoid wrist: a pilot study

To determine an optimal threshold in a simplified 3D-based volumetry of abnormal signals in rheumatoid wrists utilizing contrast and non-contrast MR data, and investigate the feasibility and reliability of this method.

MR images of bilateral hands of 15 active rheumatoid patients were assessed before and 5 months after the initiation of tocilizumab infusion protocol. The volumes of abnormal signals were measured on STIR and post-contrast fat-suppressed T1-weighted images. Three-dimensional volume rendering of the images was used for segmentation of the wrist by an MR technologist and a radiologist. Volumetric data were obtained with variable thresholding (1, 1.25, 1.5, 1.75, and 2 times the muscle signal), and were compared to clinical data and semiquantitative MR scoring (RAMRIS) of the wrist. Intra- and interobserver variability and time needed for volumetry measurements were assessed.

The volumetric data correlated favorably with clinical parameters almost throughout the pre-determined thresholds. Interval differences in volumetric data correlated favorably with those of RAMRIS when the threshold was set at more than 1.5 times the muscle signal. The repeatability index was lower than the average of the interval differences in volumetric data when the threshold was set at 1.5-1.75 for STIR data. Intra- and interobserver variability for volumetry was 0.79-0.84. The time required for volumetry was shorter than that for RAMRIS.

These results suggest that a simplified MR volumetric data acquisition may provide gross estimates of disease activity when the threshold is set properly. Such estimation can be achieved quickly by non-imaging specialists and without contrast administration.