Cartilage surface integrity influences contrast enhanced ct imaging of osteoarthritis

Purpose: To assess the utility of contrast-enhanced uCT (CE-uCT) as an imaging biomarker for preclinical assessment of osteoarthritis in a sheep model using IA injection of iodinated contrast medium and short exposure times that can be obtained in vivo. Methods: An in-vitro pilot study was conducted to assess differences in the apparent density between damaged (fibrillated/incised or shaved/superficial zone removed) and intact bovine osteochondral blocks after exposure to an anionic contrast medium (Conray 60) for 10 minutes (n=22). The sidewalls of blocks were wrapped in Parafilm™ to limit contrast contact to the cartilage surface only. Fibrillated (n=6), shaved (n=7), and intact (n=4) blocks were immersed in the Conray 60 and intact control (N=5) blocks were immersed in lactated ringer’s solution (LRS) for 10 minutes prior to uCT imaging at 45 micron resolution. Uptake of contrast was measured by calculation of the average apparent density in Hounsfield Units (HU). Twenty-four sheep were enrolled in a follow-up, ex-vivo, study. Osteoarthritis was induced by arthroscopic meniscal destabilization in n=20 sheep and n=4 remained as naive controls. The x-ray attenuation of cartilage was measured in animals with early (3 month) and late stage (12 month) OA after a 10-minute exposure to anionic contrast media (Conray 60). Concentration (ug/mg) of sulfated-glycosaminoglycan (sGAG) and histological OARSI scores for proteoglycan (PG) staining were also collected and recorded from the medial tibial plateau (MTP) of the operated leg. ANOVA was used to analyze variance in sGAG and HU between sites on the MTP or between groups. Correlation coefficients between sGAG and HU were analyzed using Pearson’s correlation coefficients and simple linear regression analysis. Spearmann-rank coefficients were used to examine the association between histological PG score and either HU or sGAG. We hypothesized apparent density (HU) would vary with respect to concentration of local sGAG, and that HU would be higher in regions where sGAG is depleted. Results: In the in-vitro CE-uCT study we observed a significant increase in apparent density of intact cartilage samples exposed to Conray 60 for 10 minutes compared to LRS controls (p=0.03). HU was also significantly higher in fibrillated (p=0.030) and shaved (p=0.0007) blocks compared to non-contrasted samples. In ex-vivo CE-uCT imaging of the MTP location accounted the variance in HU for both 3 month (p=0.09) and 12 month (p=0.0074) groups but not controls (p=0.32). Early 3 month OA sheep had elevated levels of sGAG that were significantly higher than the late phase 12 month group group (Fig 1, p =0.0032). A significant, negative correlation between sGAG concentration and HU of the same location was observed (r= -0.71, p = 0.047) in the early stage (3 month) sheep. Linear regression analysis revealed that every one unit (ug CSC/mg cartilage) increase in sGAG resulted in a reduction in HU by 8.62 units (R2 = 0.43, p< 0.05). Conclusions: Ten minutes of exposure to contrast media was sufficient to clearly outline cartilage borders due to the increase in apparent density and measured HU. Differences in cartilage HU of CE-uCT images can be used to estimate differences in local sGAG concentrations in animals with thin cartilage during early and late OA but variance in local sGAG concentration cannot exclusively explain the resulting variance in HU, suggesting that other factors may play a role such as an the integrity of the superficial cartilage zone. Elevated sGAG cartilage concentrations at three months returned to near normal levels at 12 months suggesting the hypertrophic phase of OA may be longer than expected. During this time matrix depletion as measured by CE-CT may not be a reliable imaging biomarker.