Several agents act as disease-modifying drugs in animal models of osteoarthritis (DMOADs). For a clinical trial of a DMOAD it is widely held that joint radiography, if standardized, is the best outcome measure available today. However, uncertainty exists about the reproducibility of radiographic outcome measures and the most appropriate target population for a DMOAD study. It is logical that treatment with a DMOAD whose primary effect is on biochemical/metabolic processes in articular cartilage will reduce the rate of joint space narrowing (JSN) - a surrogate for cartilage loss - in knee x-rays. Although measurements of joint space (JSW) in the conventional standing anteroposterior (AP) knee x-ray vary widely from examination to examination, it has been reported that accuracy of JSW measurements can be improved markedly with standardized positioning of the subject, but this has yet to be confirmed. Furthermore, the degree to which the severity of OA impairs reproducibility of JSW measurement is unknown.
The specific aims of this project are to: (1) confirm recent findings of very high precision (intra-technologist reproducibility) in measurement of tibiofemoral JSW obtainable with optimal positioning of the subject, computerized measurement of JSW in the digitized image and correction for image magnification, (2) determine the accuracy (i.e., inter-technologist reproducibility) of JSW measurements in x-rays of a subject taken in different radiology departments (a necessary consideration for determining the feasibility of a multi-center DMOAD trial) and (3) assess the effect of radiographic severity of OA on accuracy and precision of JSW measurements. Each of 30 subjects (10 with bilaterally normal knees [i.e., Kellgren and Lawrence Grade 0-1] and 20 with Grade II or III OA in each knee) will undergo knee x-ray in 2 of 5 participating radiology departments within a 7 day period. During the following week, each subject will return to one of the 2 facilities for a repeat study. Each knee series will consist of a conventional standing anteroposterior (AP) view in full extension and a semiflexed view with the knee positioned under fluoroscopy so that the tibial plateau is parallel to the floor. Magnification will be corrected, the image will be digitized, and JSW will be measured with a computerized system. Manual measurements made on the fully extended and semiflexed views and computerized measurements on the digitized, semiflexed view will be analyzed. Using intraclass correlation techniques, we will estimate intra- and inter-technologist reproducibility of the 3 mensural systems across all 5 facilities, and for each facility separately. Reproducibility of JSW measurement will be estimated separately for knees with Grade 0-1, II and III OA. Using multiple regression techniques we will determine the extent to which individual radiographic features of OA (e.g., osteophytes, bony sclerosis) account for within-subject variability across multiple JSW measurements.
Showing the most recent 10 out of 162 publications
