NEW JSW MEASUREMENTS INCREASE RESPONSIVENSS TO CHANGE

INTRODUCTION

Knee radiography is a low cost, convenient, and widely available modality for assessing KOA change longitudinally. Although seen on MRI, soft tissues such as cartilage and the meniscus are invisible radiographically and their change is measured indirectly as loss of radiographic JSW. This indirect association has the potential to reduce the responsiveness to change for JSW. JSW loss is likely due to a combination of cartilage and meniscus change but the level of contribution from each structure is not currently discernable from a radiograph.

OBJECTIVE

To develop and validate new measurements of JSW with improved responsiveness to change compared to the current method. We also hope this will begin to shed light on the individual contributions of cartilage and meniscus to JSW loss by systematically evaluating different JSW locations across the knee joint.

METHODS

We randomly placed all 4,796 OAI participants into either a training or testing group and selected all knees where fixed-location JSW (fJSW) was available at the x=0.15 to 0.3 (medial compartment) and x=0.7 (inner-most lateral compartment) locations at each of 6 follow-up time points (12, 24, 36, 48, 72, and 96 months). We defined a new JSW metric (JSW_New) that was a linear combination of three individual fJSW measures:

JSW_New= fJSW(x=0.25) + w₁ × fJSW(x=0.7) + w₂ × fJSW(x=x_i),

where x_i was one of 6 values in the medial compartment: 0.15, 0.175, 0.2, 0.225, 0.275 or 0.3; lower x_i values corresponded to more peripheral locations. Using the training group, we varied w₁, w₂ and x_i to achieve the maximum responsiveness, defined as the magnitude of the standardized response mean (SRM) for baseline to the follow-up time point. Once optimized, the performance was evaluated using the independent testing set and compared in the test group to the SRM found for fJSW(x=0.25), which is generally considered the most responsive fixed location JSW. We performed separate optimization and testing for the 5 different baseline KL values and 6 distinct follow-up time points.

RESULTS

Table 1 summarizes the results. There is substantial improvement in the responsiveness (magnitude of the SRM values) for all follow-up time points and KL values. We did not observe a consistent pattern for the x_i values other than the absence of x=0.15 (most peripheral) as an optimal value. w₁ was generally negative for KL4 knees suggesting that JSW_New may be capturing pseudo-widening (seesaw effect) or possibly medial compartment meniscus extrusion for these knees. w₂, the weight factor for the other medial compartment locations, was consistently positive although no discernable dependence on KL or follow-up time point was observed. Positive w₂ is consistent the understanding that KOA is generally a medial compartment disease.

CONCLUSION

We report more responsive JSW metrics that have the potential to improve the utility of radiographs for clinical trials and other studies of KOA. The results suggest that this measurement may capture additional information related to the underlying cartilage and meniscal changes. However, further work incorporating MRI data will be necessary to better understand the patterns of locations and weights and the full implications of these results.