Background/Purpose: Thumb base osteoarthritis (OA) is characterized by structural damage, most commonly evaluated by radiography. However radiography is insensitive. Magnetic resonance (MR) imaging could be a valuable alternative, however a standardized scoring method for thumb base MR images did not exist until recently OMERACT developed the thumb base OA MRI scoring system (TOMS). Our aim was to investigate the validity of the new TOMS by comparing TOMS scores with radiographic scores.

Methods: Two hundred patients (83.5% women, median age 60.5 years) with primary hand OA (90.5% fulfilling ACR classification criteria) from the rheumatology outpatient clinic, who had dorsopalmar radiographs and MR scans of the right thumb base, were studied. T1- and fat suppressed T2-turbo spin weighted sequences were obtained on a 1.5 Tesla scanner. Radiographs of the first carpometacarpal (CMC1) and scaphotrapeziotrapezoid (STT) joints were scored using the OARSI atlas (osteophytes and joint space narrowing [JSN] in CMC1: 0-3 and STT: absent/present) in consensus by two readers with good intra-reader reliability. MR images were scored using TOMS (osteophytes in CMC1: 0-6 and STT: 0-9; cartilage space loss [CSL] for both joints: 0-3) by two readers, with good intra- and inter-reader reliability. For analysis we used the average of both readers. Readers were blinded for clinical and other imaging data. To study validity, the distribution of the TOMS scores for osteophytes and CSL were described stratified for the different radiographic stages for osteophytes and JSN, respectively.

Results: On MR images osteophytes were detected in the vast majority of thumb bases (CMC1 n=172; STT n=102). The score of TOMS increased with more severe radiographic stages (see figures). However, the number of patients without any osteophytes in both CMC1 and STT was considerably lower for TOMS (n=19) than for the OARSI (n=105) scoring. A similar difference was apparent for absence of CSL (n=82) versus JSN (n=107) in both CMC1 and STT. Patients with isolated STT osteophytes were quite rare for both TOMS (n=9) and the OARSI (n=1) scoring. The most prominent discrepancy between TOMS and OARSI sensitivity was found for osteophytes: an additional 170 joints (CMC1 n=79; STT n=91) were found positive with TOMS, while only 1 OARSI-positive CMC1 scored negative with TOMS.

Conclusion: Scores of OA features assessed on MR images by TOMS were correlated with radiographic scores, indicating good validity of the TOMS. Furthermore, the frequencies of positive features assessed on MR images were higher compared to those on radiographs, suggesting high sensitivity for the TOMS.