Philipp Lang MD, MBA, PRESENTER: Nothing to Disclose

Abstract: HTML Purpose: To determine the reproducibility of radiologists' visual scoring of cartilage loss on MRI scans. Methods and Materials: MR images were obtained in 28 patients with osteoarthritis of the knee (Kellgren-Lawrence grade I and II) participating in a clinical trial evaluating the efficacy of a chondroregenerative drug. MR images were obtained with a 1.5 T scanner (GE Signa). Standard sagittal fat-suppressed FSE (TR/TE in msec 4000/13 and 4000/39, ETL 8, 4-mm slice thickness, 1-mm intersection gap, matrix 256x192, field of view (FOV) 120 mm) and sagittal fat-suppressed 3D SPGR images (60/5/40° flip angle, 1 NA (number of signal averages), 1.5-mm slice thickness, matrix of 256x192, FOV 120mm) were acquired in each subject. Baseline MRI scans were read twice by two experienced musculoskeletal radiologists using our previously published scoring system (Cartilage loss grade 0 [normal] to 6 [full thickness loss], cartilage loss size A to E [A£1cm2; B£2cm2 & >1cm2; C£3cm2 & >2cm2; D£4cm2 & >3cm2; E >4cm2]). Statistical intra- and inter-reader reliability analyses were conducted based on both (1) percent agreement and (2) weighted kappa (result for kappa strength of agreement defined as: < 0.00 poor, 0.00 - 0.20 slight, 0.21 - 0.40 fair, 0.41 - 0.60 moderate, 0.61 - 0.80 substantial, 0.81 - 1.00 almost perfect). Results: (1) Ranges of the percent agreement: Grade: intra-reader percent agreement ranged from 67.9 - 96.1%; inter-reader percent agreement ranged from 37.1 - 84.0%. Size: intra-reader percent agreement ranged from 57.7 - 82.2%; inter-reader percent agreement ranged from 44.4 - 85.2%. (2) Ranges of kappa: For grade, intra-reader kappa ranged from 0.52 - 0.90; inter-reader kappa ranged from 0.06 to 0.57; For size, intra-reader kappa ranged from 0.36 to 0.62; inter-reader kappa ranged from 0.01 to 0.61. Conclusion: Kappa value and percent agreement were typically lower for size than for grade. Grade, i.e. the depth of a cartilage defect, is best assessed on the slice passing through the deepest part of a cartilage defect. The accuracy of the assessment of grade is largely dependent on the in-plane resolution which is high for both 2D FSE and 3D SPGR. Size is an areal measurement and requires assessment of several consecutive slices. Thus, the accuracy and reproducibility of size measurements will deteriorate with increasing slice thickness. The moderate kappa scores observed for grade and size emphasize the need for novel 3D imaging techniques with improved contrast and, preferably, near isotropic resolution.       Questions about this event email: pklang@partners.org

Lang MD, MBA, P, MRI of Articular Cartilage: Reproducibility of Visual Readings of Cartilage Loss Using 2D FSE and 3D SPGR Sequences.  Radiological Society of North America 2003 Scientific Assembly and Annual Meeting, November 30 - December 5, 2003 ,Chicago IL. http://archive.rsna.org/2003/3105570.html