Assessment of cartilage changes over time in knee osteoarthritis disease-modifying osteoarthritis drug trials using semiquantitative and quantitative methods: pros and cons.

OBJECTIVE: To evaluate the impact of 2 magnetic resonance imaging (MRI) sequences on cartilage defect assessment in knee osteoarthritis (OA) patients and the sensitivity to change over time comparing cartilage defect (semiquantitative) with cartilage volume loss (quantitative) methods.

METHODS: Gradient-echo (GRE) and intermediate-weighted fast spin-echo (IW-FSE) sequences were compared. Knee OA MRIs were from two 2-year studies (cohort 1, n = 55; cohort 2, n = 143). For both cohorts, a GRE sequence was used and patients in cohort 1 underwent an additional IW-FSE sequence. Cohort 2 included patients from a previous trial. Cartilage defects and cartilage volume were evaluated.

RESULTS: The cartilage defect assessment provided consistently significantly higher scores in IW-FSE than in GRE sequences at baseline and 2 years. However, there was no difference in the change at 2 years between the sequences. The standardized response mean (SRM) for change did not show a difference between the 2 sequences, but was consistently higher (2-2.5-fold) for the quantitative method. The cartilage defect score change between the 2 treatment groups revealed a trend toward significance only in the medial tibial plateau, whereas the change in cartilage volume loss demonstrated a significant difference in the global knee, global femur, lateral femur, and lateral compartment. The SRMs for the treatment groups combined were markedly higher for cartilage volume loss than for the defect scoring by 4.3- to 6.0-fold.

CONCLUSION: The direct comparison between GRE and IW-FSE sequences did not suggest superior sensitivity to cartilage defect change over time of one sequence over the other. Interestingly, the quantitative cartilage volume assessment was more sensitive than the semiquantitative scoring in the detection of treatment effect on OA cartilage changes.