Mylène Jansen

Comparing radiographic JSW and MRI cartilage thickness using CT 247 12 the exact same imaging technique. Results of such future studies could help to better relate results obtained from MRI scans and radiographs to monitor OA progression or treatment response. An important consideration in using weight-bearing CT or MRI is that using such approaches need thorough concern of the relative contribution of weight and cartilage deformability. Also, the actual weight-bearing relative to the contra-lateral leg in case of uneven load distribution as well as preacquisition weight-bearing or exercise is a parameter to consider in such a study. 37 A limitation of our study is the relatively small sample size, as only 14 of the originally 20 complete data sets were available. As a sensitivity analysis, the 2 patients that were excluded because of MRI motion artifacts were included in the evaluation of radiographic JSW, 2D CT JSW and 3D CT JSW. The significance of the correlations between these 3 techniques for these 16 patients did not change compared to the (for all images complete) dataset of 14 patients, neither for absolute (cross-sectional) values nor for changes over time. Also, scatterplot matrices of all calculated correlations were included, because p- values may be less conclusive in this small number of patients. Clearly the scatterplot matrices support the conclusions based on the Pearson R and p- values. Irrespectively, the present study is a post-hoc analysis and exploratory. More research with larger data sets, preferable using weight-bearing CT or MRI as additional variables, would validate the conclusion. Another limitation of our study is that knee flexion is not taken into account. The weight- bearing radiographs are performed under slight flexion of the knee (7 – 10°). MRI and CT scans are not performed under a specific angle, but normally the leg is extended for as much as is allowed by, for example, a patient’s possible extension limitation or the hardware setup. Although the 3D imaging techniques provide a mean or median surface value, the 2D rendering of the 3D CT has a potential knee flexion angle difference as compared to the plain radiograph. This difference might have influenced the correlation between both techniques and the effect of different knee flexion could be included in future research as well. In conclusion, the cause of the generally weak correlation between changes in radiographic JSW and MRI cartilage thickness appears to primarily be the difference in weight-bearing conditions during imaging, and less so the difference in measuring bone-to-bone distance versus cartilage thickness directly. Further research on the effects of weight-bearing on cartilage thickness measurements is warranted and might provide an indirect measure for cartilage deformability in case of quantitative measurements, in addition to the measured thickness.