Background/Purpose: Knee joint replacement (KR) is a cost-effective procedure with good long-term outcomes. However, at present there is no clear consensus on indications for KR. Imaging biomarkers capable of predicting KR therefore are urgently needed and may be helpful in clinical studies and trials that utilize KR as an outcome. Specific MRI-based quantitative measures of cartilage morphology have recently been described to be associated with an increased risk of KR. The aim of the present study was to determine whether longitudinal changes in semi-quantitative MRI-based measures of cartilage differ between knees undergoing KR and control knees not undergoing KR. We used a compartment-based analysis (i.e., medial tibio-femoral (MTF), lateral tibio-femora (LTF) or patello-femoral (PF)0.

Methods: We studied 127 knees from OAI participants (age 46-81 years) that underwent KR before the 60 month visit at two time points prior to KR, (e.g. for a KR reported at the 48 month (M) visit, T0 = 36M and T-1 = 24M).127 control knees that did not undergo KR were matched for KL grade (0/1, 2, 3, 4), gender, and age (+/- 5y) and were assessed at the same T0 and T-1 visits. 3 T MRIs were read for cartilage morphology using the semiquantitative MOAKS system in 14 articular subregions, which scores cartilage from 0 to 3 in regard to proportion of the area of the subregion involved and from 0 to 3 in regard to percentage of subregion involved by full thickness loss. Within-grade coding was used to score definite visual changes that do not fulfill criteria of a full grade change. Conditional logistic regression was applied to assess compartmental changes in cartilage morphology from T-1 to T0 with respect to risk for KR following T0.

Results: Subjects were on average 65.7 years old (SD ± 8.8), predominantly female (59.8%) and overweight (mean BMI 29.7 SD ± 4.82). The K/L distribution was: KL0: 8 (3.15%) KL1: 8 (3.15%), KL2: 57(22.44%), KL3:146 87(34.25%) and KL4: 94 (37.01%).

Development of full thickness cartilage damage and full grade worsening in the LTF, but not MTF compartment from T-1 to T0 was associated with greater odds for KR (OR 2.8, 95% CI 1.01-7.78 and 4.5 95% CI 1.52-13.30, respectively). More than 3 subregions exhibiting worsening in the MTF compartment was associated with increased odds for KR (OR 14.5, 95% CI 1.90-110.93), while this was not observed for the LTF or PF compartments.

Conclusion: Development of full thickness cartilage damage and worsening in the LTF compartment as well as multiple subregions showing worsening cartilage damage in the MTF compartment increase odds for TKR. Presence and change of these imaging biomarkers may be important prognostic markers when KR is used as a long-term outcome.