Abstract
Objective. To develop and validate a knowledge transfer (KT) module aimed at enhancing feasibility and reliability of semiquantitative assessment of bone marrow lesions (BML) and synovitis-effusion using the Hip Inflammation Magnetic Resonance Imaging Scoring System (HIMRISS).
Methods. Three radiologists naive to the HIMRISS method reviewed the manuscript describing the method and then scored MRI scans from 16 patients with hip OA obtained at baseline and 8 weeks after intraarticular injection of corticosteroid. Readers then reviewed a KT module comprising an instructional presentation and 8 reference DICOM (digital imaging and communications in medicine) cases scored by 3 readers with expertise in the HIMRISS method, and then used electronic overlay software to score scans from 23 patients with OA. The same format was followed with a second group of 3 readers naive to HIMRISS using a KT module revised to incorporate the overlay with a Web-based DICOM viewer to enhance feasibility. Interobserver reliability was assessed with the intraclass correlation coefficient (ICC).
Results. In both exercises, reliability for baseline scores was excellent for femoral BML, very good for acetabular BML, and good for synovitis-effusion (overall ICC = 0.91, 0.89, 0.62, respectively) even without prior calibration using the KT module. However, reliability for detecting change was substantially worse than for expert readers, especially for acetabular BML and synovitis-effusion (overall ICC = 0.59 vs 0.19, and 0.42 vs 0.25, respectively). Reliability improved for detection of change in these lesions, especially after reader calibration with the revised KT module.
Conclusion. Development and validation of a systematic method for KT may enhance external validation of certain imaging instruments.
