Article Text

SAT0335 Semi-quantitative assessment of bone marrow edema and synovitis-effusion in osteoarthritis with the knee inflammation MRI scoring system (KIMRISS): A target lesion based methodology
  1. W. Maksymowych1,
  2. U. Weber2,
  3. M. Pianta1,
  4. R. Lambert1
  1. 1University of Alberta, Edmonton, Canada
  2. 2Balgrist University Hospital, Zurich, Switzerland


Background Current MRI scoring methods for assessment of acute lesions such as bone marrow edema (BME) in the knee of patients with osteoarthritis rely on a complex subdivision of the knee into 15 subregions and then a further estimation of the proportion of subregion with BME1. This limits feasibility for widespread adoption. Scoring of synovitis-effusion (S-E) is based on a restricted grading scheme assessing the whole joint (0= none, 3 = large) which limits responsiveness, especially for interventions that might target inflammation.

Objectives To develop and conduct preliminary validation of an MRI method (KIMRISS) for direct semi-quantitative assessment of acute lesions, BME and S-E, that focuses on detection of change.

Methods Assessment of BME is based on assessment of coronal and sagittal images for medial/lateral knee compartments and axial/sagittal images for patella-femoral compartment using a fluid-sensitive MRI sequence (STIR, T2 FatSat). Size of a BME lesion is defined according to the largest continuous increase in signal assessed in all dimensions and number of slices in which the increased signal can be detected (small = <1cm in all dimensions on ≤2 slices; moderate = >1cm but NOT >2 cm in ≥2 dimensions; large = >2cm in ≥2 dimensions). A weighting is applied to change in BME size (1.5x and 2x for moderate and large lesions, respectively). Size of S-E is assessed in each of 4 compartments (medial and lateral patellar recess, suprapatellar, semimembranosus bursa) according to a 0-4 grading scheme and a weighting is applied for change in S-E size (1.5x and 2x for grade 3 and 4 lesions, respectively). MRI scans were performed on the knee joints of 15 patients enrolled into an open label trial of an anti-TNF agent in subjects with persistent pain due to knee osteoarthritis and clinical evidence of effusion who had failed conventional therapy. Scans were performed at baseline and 12 weeks and independently reviewed by 3 readers blinded to timepoint. Reliability of change scores was assessed by intraclass correlation coefficient (ICC) and responsiveness by standardized response mean (SRM). We assessed correlations with WOMAC pain, patient global, and target joint clinical effusion score.

Results Reliability of detection of change in KIMRISS BME (ICC for 3 reader pairs =0.71. 0.73. 0.75), KIMRISS S-E (ICC for 3 reader pairs =0.78, 0.82, 0.86), and Total KIMRISS (ICC for 3 reader pairs =0.77, 0.81, 0.89) was very good with substantial responsiveness after 12 weeks of treatment (Table). Improvement in Total KIMRISS score was observed in 12 patients although change in either the Total KIMRISS score or KIMRISS BME did not significantly correlate with change in WOMAC pain or patient global. KIMRISS S-E did not correlate with target joint effusion score.

Conclusions The KIMRISS methodology for MRI-based semi-quantitative assessment of acute lesions in knee joints is responsive and is capable of reliably detecting change. It merits further validation in inflammatory knee joint disorders.

  1. Hunter et al. Ostearthritis Cartilage 2011; 19: 990

Disclosure of Interest None Declared

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.