Objective

Minimum radiographic joint space width (mJSW) represents the Food and Drug Administration (FDA) standard for demonstrating structural therapeutic benefits for knee osteoarthritis (KOA), but only shows moderate responsiveness (sensitivity to change). We directly compare the responsiveness of magnetic resonance imaging (MRI)-based cartilage thickness and JSW measures from fixed-flexion radiography (FFR) and explore the correlation of region-matched changes between both methods.

Methods

Nine hundred and sixty-seven knees of Osteoarthritis Initiative participants with radiographic KOA were studied: 445 over 1 year with coronal FLASH MRI and FFR, and 375/522 over 1/2 years with sagittal DESS MRI and FFR. Standardized response means (SRM) of cartilage thickness and mJSW were compared using the sign-test.

Results

With FLASH MRI, SRM was -0.28 for medial femorotibial compartment (MFTC) cartilage loss vs -0.15 for mJSW, and -0.32 vs -0.22 for the most sensitive MRI subregion (central MFTC) vs the most sensitive fixed-location JSW(x = 0.25). With DESS MRI, 1-year SRM was -0.34 for MFTC vs -0.22 for mJSW and -0.44 vs -0.28 for central MFTC vs JSW(x = 0.225). Over 2 years, the SRM was significantly greater for MFTC than for mJSW (-0.43 vs -0.31, P = 0.017) and for central MFTC than for JSW(x = 0.225) (-0.51 vs -0.44, P < 0.001). Correlations between changes in spatially matched MRI subregions and fixed-location JSW were not consistently higher (r = 0.10-0.51) than those between non-matched locations (r = 0.15-0.50).

Conclusions

MRI displays greater responsiveness in KOA than JSW FFR-based JSW, with the greatest SRM observed in the central medial femorotibial compartment. Fixed-location radiographic measures appear not capable of determining the spatial distribution of femorotibial cartilage loss.

Objective

To determine the predictive value of unicompartimental joint space narrowing (JSN) for MRI-based cartilage thickness loss in the narrowed and the non-narrowed femorotibial compartment.

Methods

922 knees from 922 Osteoarthritis Initiative (OAI) participants (62.2 ± 9.0 years, 61% females) with radiographic OA (158 without JSN [noJSN], 175 with lateral JSN [latJSN], 589 with medial JSN [medJSN]) were analyzed using 3 T MRI. One-year cartilage thickness change was determined in the lateral (LFTC) and medial femorotibial compartment (MFTC), and in femorotibial subregions. The probability of subsequent cartilage loss was calculated using predefined thresholds. The predictive value of JSN for the probability and magnitude of cartilage loss was compared between latJSN, medJSN and noJSN knees using Fisher's exact and Mann-Whitney-U tests.

Results

The probability of cartilage loss was greater in the narrowed compartment of latJSN/medJSN knees (34.9%/32.4%) than in noJSN knees (13.3%/12.7%, P ≤ 6.4 × 10(-6)) and so was the magnitude of cartilage thickness change (P ≤ 8.2 × 10(-6)). No significant differences were observed between the narrowed compartments of latJSN vs. medJSN knees (probability: P = 0.58, magnitude: P = 0.19) or between the non-narrowed compartment of latJSN/medJSN vs. noJSN knees (probability: P ≥ 0.35, magnitude: P = ≥0.23). These results were confirmed by the location-independent ordered value (OV) analyses of femorotibial subregions.

Conclusion

The predictive value of latJSN for lateral compartment cartilage loss was comparable to that of medJSN for medial compartment cartilage loss, whereas cartilage loss in the non-narrowed compartment was similar to that in noJSN knees. These findings provide important clues to predicting progression of knee OA, and in tailoring inclusion criteria for clinical trials.

Objective

To compare unbiased estimates of short- vs long-term cartilage loss in osteoarthritic knees.

Method

441 knees [216 Kellgren Lawrence (KL) grade 2, 225 KL grade 3] from participants of the Osteoarthritis Initiative were studied over a 4-year period. Femorotibial cartilage thickness was determined using 3 T double echo steady state magnetic resonance imaging, the readers being blinded to time points. Because common measurement time points bias correlations, short-term change (year-1 to year-2: Y1 → Y2) was compared with long-term change (baseline to year-4: BL → Y4), and initial (BL → Y1) with subsequent (Y2 → Y4) observation periods.

Results

The mean femorotibial cartilage thickness change (standardized response mean) was -1.2%/-0.8% (-0.42/-0.28) over 1 (BL → Y1/Y1 → Y2), -2.1%/-2.5% (-0.56/-0.55) over 2 (BL → Y2/Y2 → Y4), -3.3% (-0.63) over 3 (Y1 → Y4), and -4.5% (-0.78) over 4 years. Spearman correlations were 0.33 for Y1 → Y2 vs BL → Y4, and 0.17 for BL → Y1 vs Y2 → Y4 change. Percent agreement between knees showing progression during Y1 → Y2 vs BL → Y4 was 59%, and 64% for BL → Y1 vs Y2 → Y4. The area under the receiver operating characteristic curve was 0.66 for using Y1 → Y2 to predict BL → Y4, and 0.59 for using BL → Y1 to predict Y2 → Y4 change.

Conclusion

Weak to moderate correlations and agreement were observed between individual short- vs long-term cartilage loss, and between initial and subsequent observation periods. Hence, longer observation periods are recommended to achieve robust results on cartilage loss in individual knees. At cohort and subcohort level (e.g., KLG3 vs KLG2 knees), the mean cartilage loss increased almost linearly with the length of the observation period and was constant throughout the study.