Pos1364 a comparison of semi-quantitative and quantitative cartilage loss assessments: data from the osteoarthritis initiative

Background In Disease Modifying Osteoarthritis Drug (DMOAD) trials, cartilage loss may be assessed semi-quantitatively or quantitatively, although the latter is higher cost and more technically challenging. The MRI Osteoarthritis Knee Score (MOAKS) is the most commonly used semi-quantitative knee OA scoring system that assesses thickness of MOAKS cartilage morphology loss (tMCM) and denudation (dMCM) on the articulating subregions of femur and tibia. Objectives This study aimed to compare the cross-sectional relationship and longitudinal responsiveness of MOAKS with the more accurate quantitative cartilage assessment. Methods Images and MOAKS scores from 297 participants with radiographic progression (groups 1 and 2) from the OAI FNIH sub-cohort were included. Quantitative cartilage thickness was measured using Active Appearance Models (AAMs). To facilitate direct comparison with MOAKS, novel quantitative measures of cartilage loss were matched to MOAKS regions (Q-MOAKS). Mean normative cartilage thickness was computed for each subregion (Figure 1) using FNIH controls (group 4). Q-MOAKS thickness loss score (tQCM) was based on the proportion of cartilage thickness over a subregion that was <95% normative thickness and denudation score (dQCM) was based on <5% normative thickness. Q-MOAKS area proportions were categorised into scores as for MOAKS (0: none, 1: 0-10%, 2: 10-75% and 3: >75%). Quantitative cartilage thickness (ThCtAB) was also measured in the central medial femur (cMF) and tibia (cMT). We compared MOAKS against Q-MOAKS and ThCtAB. Cross-sectional relationships between measures were assessed using Spearman’s rank correlation. Responsiveness was assessed at 1 and 2 years using bootstrapped standardised response means (SRM). Figure 1: Cartilage regions definitions consisting of the medial and lateral central and posterior femur (cMF, cLF, pMF, pLF) subregions and the medial and lateral tibial anterior, central, posterior (aMT, cMT, pMT, aLT, cLT, pLT) subregions. Results Cross-sectionally, there was moderate correlation between MOAKS and Q-MOAKS denudation in the central medial femur (cMF r = 0.42, (95%CI: 0.32, 0.51)) and tibia (cMT r = 0.51, (0.42, 0.59)). There was a poor correlation between MOAKS and Q-MOAKS thickness loss and denudation scores in all other regions. In the tibia (cMT), 61% (96/159) of knees with thickness loss tMCM = 2 (the 10-75% score) were also tQMC = 2 and 66% of denudation dMCM = 2 were also dQCM = 2. In the femur (cMF), the figures were 56% and 23%. MOAKS tMCM and dMCM were less responsive than Q-MOAKS tQCM and dQCM in most subregions (selected SRMs presented in Table 1). Q-MOAKS tMCM in pMT demonstrated the most responsiveness for all the scores (SRM=+0.76 vs SRM = +0.20 for MOAKS tMCM). Quantitative cartilage thickness (ThCtAB) measures were most responsive. In the cMF region, the SRM= -0.46 (95% CI: -0.56, -0.33) at 1-year and 0.82 (-0.94, -0.72) at 2-years while the cMT SRM = -0.4 (-0.5, -0.28) at 1-year and -0.71 (-0.80, -0.60) at 2-years. Table 1. Responsiveness of cartilage scores at follow up. Values are SRM (95% CI). YEAR ONE YEAR TWO AREA tMCM tQCM dMCM dQCM tMCM tQCM dMCM dQCM cMF +0.28 (0.13,0.41) +0.32 (0.20,0.43) +0.07 (-0.05,0.18) +0.20 (0.11,0.28) +0.46 (0.32,0.57) +0.53 (0.40,0.63) +0.37 (0.26,0.48) +0.44 (0.38,0.51) cLF +0.07 (-0.07.0.17) +0.20 (0.09,0.33) +0.02 (-0.09,0.14) +0.14 (0.05,0.23) +0.13 (0.02,0.20) +0.24 (0.12,0.36) +0.11 (-0.03,0.19) +0.30 (0.24,0.36) cMT +0.32 (0.27,0.37) +0.36 (0.24,0.47) +0.23 (0.11,0.32) +0.25 (0.13,0.33) +0.45 (0.37,0.52) +0.61 (0.49,0.72) +0.42 (0.29,0.53) +0.44 (0.36,0.51) cLT +0.18 (0.12,0.24) 0.22 (0.10,0.34) +0.11 (-0.03,0.18) +0.15 (0.04,0.24) +0.23 (0.17,0.29) +0.14 (0.02,0.26) +0.17 (0.08,0.23) +0.29 (0.20,0.37) Conclusion Quantitative measures or derived Q-MOAKS scores were more responsive than MOAKS. MOAKS appears to have better concordance with quantitative cartilage denudation than thickness loss, which might be explained by diagnostic accuracy for full thickness lesions being better than earlier grade lesions. REFERENCES: NIL. Acknowledgements: NIL. Disclosure of Interests Aaron Ray: None declared, Bright Dube: None declared, Michael A Bowes: None declared, Alan Brett: None declared, Emma Rowbotham: None declared, Philip G Conaghan Speakers bureau: AbbVie, Novartis, Consultant of: AbbVie, AstraZeneca, Biosplice, BMS, Eli Lilly, Galapagos, Genascence, GSK, Merck, Novartis, Pfizer, Regeneron, Stryker, and UCB.