[OP0036] INTER-READER RELIABILITY OF A PROPOSED PRELIMINARY MAGNETIC RESONANCE IMAGING (MRI) SCORING SYSTEM IN HAND OSTEOARTHRITIS (HOA)
I.K. Haugen1, S. Lillegraven1, B. Slatkowsky-Christensen1, E. Haavardsholm1, T.K. Kvien1, D. van der Heijde1 2, S. Sesseng3, P. Bøyesen1. 1Dept. of Rheumatology, Diakonhjemmet Hospital, Oslo, Norway; 2Dept. of Rheumatology, Leiden University Medical Center, Leiden, Netherlands; 3Dept. of Radiology, Diakonhjemmet Hospital, Oslo, Norway
Objectives: To examine the inter-reader reliability of the proposed features (subscales) in the MRI scoring system in HOA.
Methods: The exercise was performed using MRI scans from 10 HOA patients (9 women) with mean (SD) age of 64.8 (4.8) years from the Oslo HOA cohort. The patients were selected from quintiles of radiographic severity. The dominant hand was anchored in a cylindrical coil (diameter 100 mm), and the distal (DIP) and proximal interphalangal (PIP) joints of the 2nd-5th fingers were examined using a high-field extremity MRI unit (1.0 T). Coronal, sagittal and axial T1-weighted fat-suppressed pre-and postgadolinium images were reproduced from a 3D dual echo Dixon technique (TR 20 ms, TE 5 ms, 1 mm slice thickness), in addition to coronal and axial Short T1 Inversion Recovery (STIR) images (TR 2850 and 3150 ms, TE 16.3 and 21 ms, 2 and 3 mm slice thickness, respectively). A training session was arranged prior to the study. The images were read by three assessors (IKH, SL, PB) blinded for patient characteristics. Reading followed the proposed MRI scoring system of DIP and PIP joints: Synovitis grade 0-3, flexor tenosynovitis 0-3, erosions 0-10 per bone, cysts 0-1 per bone, osteophytes (OP) 0-3 per bone, joint space narrowing (JSN) 0-3, malalignment (0-1 in frontal and sagittal plane), bone marrow lesions (BML) 0-3 per bone and collateral ligament presence 0-1 bilaterally, thickening 0-1 bilaterally and BML at insertion sites 0-1 bilaterally per bone. The inter-reader reliability of the individual subscales was assessed by intra-class correlation coefficients (ICC) (two way random, absolute agreement, single measures).
Results: Table 1 shows mean (SD) scores between the three readers and ICC values (95% confidence interval (CI)) for all three readers. The range of each feature is reported in brackets. The inter-reader reliability was excellent for OP and good for all other features with exception of flexor tenosynovitis, collateral thickening and BML at collateral insertion sites.
|Mean (SD)||ICC (95% CI)|
|Synovitis [0-24]||9.0 (3.1)||0.64 (0.26-0.88)|
|Flexor tenosynovitis [0-24]||3.1 (2.1)||0.34 (0.00-0.72)|
|Erosions [0-160]||7.4 (3.7)||0.74 (0.40-0.92)|
|Cysts [0-16]||1.1 (1.1)||0.64 (0.30-0.88)|
|OP [0-48]||20.1 (10.3)||0.92 (0.79-0.98)|
|JSN [0-24]||7.0 (3.2)||0.64 (0.10-0.90)|
|Malalignment [0-16]||0.7 (0.7)||0.65 (0.31-0.89)|
|BML [0-48]||3.4 (2.4)||0.71 (0.40-0.91)|
|Collateral ligament presence [0-16]||10.4 (3.0)||0.52 (0.05-0.84)|
|Collateral ligament thickening [0-16]||4.8 (2.3)||0.28 (-0.03-0.68)|
|BML at insertion of collateral ligament [0-32]||1.7 (1.1)||0.33 (0.00-0.71)|