48 patients with knee osteoarthritis were randomized to receive treatment with Structum (chondroitin sulfate) or a placebo. Over 48 weeks, the Structum group demonstrated an increase in total cartilage volume, whereas the control group demonstrated a decrease. However, this difference was not statistically significant. At 4 weeks, scores for pain, while in motion, were significantly lower in the Structum group than in the control group. The two groups did not differ with respect to any other outcomes. There was no significant correlation between clinical scores and MRI outcomes. In addition, MRI was deemed as a useful tool for clinical trials, as it was associated with good intra- and inter-reader reliability.