M Lu1, Z Chen2, W Han3, Z Zhu4, X Jin3, D J Hunter5, C Ding6. 1. Menzies Institute for Medical Research, University of Tasmania, Hobart, Tasmania, Australia; Department of Orthopaedics, 1st Affiliated Hospital, Anhui Medical University, Hefei, Anhui, China; Arthritis Research Institute, 1st Affiliated Hospital, Anhui Medical University, Hefei, Anhui, China. 2. Menzies Institute for Medical Research, University of Tasmania, Hobart, Tasmania, Australia; School of Engineering, Southeast University, Nanjing, China. 3. Menzies Institute for Medical Research, University of Tasmania, Hobart, Tasmania, Australia. 4. Menzies Institute for Medical Research, University of Tasmania, Hobart, Tasmania, Australia; Arthritis Research Institute, 1st Affiliated Hospital, Anhui Medical University, Hefei, Anhui, China. 5. Institute of Bone and Joint Research, Kolling Institute, University of Sydney; Rheumatology Department, Royal North Shore Hospital, Sydney, NSW, Australia. 6. Menzies Institute for Medical Research, University of Tasmania, Hobart, Tasmania, Australia; Arthritis Research Institute, 1st Affiliated Hospital, Anhui Medical University, Hefei, Anhui, China; Institute of Bone and Joint Research, Kolling Institute, University of Sydney; Rheumatology Department, Royal North Shore Hospital, Sydney, NSW, Australia. Electronic address: Changhai.Ding@utas.edu.au.
Abstract
PURPOSE: To assess reliability and validity of a semi-automated quantitative method to measure infrapatellar fat pad (IPFP) signal intensity in patients with knee osteoarthritis (OA). METHODS: Hundred patients with knee OA were selected. Sagittal planes of fat-saturated T2-weighted images obtained on 1.5-T magnetic resonance imaging (MRI) were utilized to assess IPFP signal intensity using MATLAB. Knee structural abnormalities including cartilage defects, bone marrow lesions (BML) and radiographic OA (ROA) were evaluated. Clinical construct validity and concurrent validity were examined through describing associations of IPFP measurements with knee structural abnormalities and a semi-quantitative scoring method, respectively. The reliability was examined by calculating the intra- and inter-observer correlation coefficients. RESULTS: Significantly positive associations were found between standard deviation of IPFP signal intensity [sDev (IPFP)], clustering factor (H) and all knee structural abnormalities. The volume of high signal intensity regions [Volume (H)] and the ratio of Volume (H) to volume of whole IPFP [Percentage (H)] were positively associated with cartilage defects and ROA, but not with BMLs. The median value [Median (H)] and upper quartile value [UQ (H)] of high signal intensity were only significantly associated with quartiles of cartilage defect score. Significant correlations were found between all quantitative measurements and semi-quantitative scores (All P < 0.001). Intraclass and interclass correlation coefficients for these quantitative measures were high (>0.90). CONCLUSIONS: A novel and efficient method to segment IPFP and calculate its signal intensity on T2-weighted MRI images is documented. This method is reproducible, and has concurrent and clinical construct validity, but its predictive validity needs to be examined by future longitudinal studies.
PURPOSE: To assess reliability and validity of a semi-automated quantitative method to measure infrapatellar fat pad (IPFP) signal intensity in patients with knee osteoarthritis (OA). METHODS: Hundred patients with knee OA were selected. Sagittal planes of fat-saturated T2-weighted images obtained on 1.5-T magnetic resonance imaging (MRI) were utilized to assess IPFP signal intensity using MATLAB. Knee structural abnormalities including cartilage defects, bone marrow lesions (BML) and radiographic OA (ROA) were evaluated. Clinical construct validity and concurrent validity were examined through describing associations of IPFP measurements with knee structural abnormalities and a semi-quantitative scoring method, respectively. The reliability was examined by calculating the intra- and inter-observer correlation coefficients. RESULTS: Significantly positive associations were found between standard deviation of IPFP signal intensity [sDev (IPFP)], clustering factor (H) and all knee structural abnormalities. The volume of high signal intensity regions [Volume (H)] and the ratio of Volume (H) to volume of whole IPFP [Percentage (H)] were positively associated with cartilage defects and ROA, but not with BMLs. The median value [Median (H)] and upper quartile value [UQ (H)] of high signal intensity were only significantly associated with quartiles of cartilage defect score. Significant correlations were found between all quantitative measurements and semi-quantitative scores (All P < 0.001). Intraclass and interclass correlation coefficients for these quantitative measures were high (>0.90). CONCLUSIONS: A novel and efficient method to segment IPFP and calculate its signal intensity on T2-weighted MRI images is documented. This method is reproducible, and has concurrent and clinical construct validity, but its predictive validity needs to be examined by future longitudinal studies.
Authors: A Ruhdorfer; F Haniel; T Petersohn; J Dörrenberg; W Wirth; T Dannhauer; D J Hunter; F Eckstein Journal: Osteoarthritis Cartilage Date: 2017-02-12 Impact factor: 6.576
Authors: Kang Wang; Changhai Ding; Michael J Hannon; Zhongshan Chen; C Kent Kwoh; David J Hunter Journal: Arthritis Care Res (Hoboken) Date: 2019-01 Impact factor: 4.794