Differentiating Healthy Cartilage and Damaged Cartilage Using Magnetic Resonance Images in a Quantitative Manner

This paper presents a study that performs a statistical analysis of signal intensities of the cartilage using magnetic resonance images. The aim of the study is to investigate whether it is possible to differentiate cartilage that is normal and cartilage that has damage/lesions in a quantitative manner. Because damaged cartilage tends to have abnormally high signal intensities than that of normal cartilage in fast spin echo proton density weighted (PD) images, we hypothesize that there is a relationship between the signal intensities of the cartilage and the size of the damaged cartilage presence. Twelve MR data sets with different degrees of cartilage damage and five data sets of normal cartilage were used in this study. Femoral articular cartilage was manually segmented using PD images and the MR signal intensities of the cartilage were analyzed. Results show that there is a linear relationship between the difference in mean and median of the cartilage signals (mean-median) and the percentage of damaged cartilage presence (R² = 0.799, p <; 0.01). The results also showed that when the cartilage has minor or no damage, the sign of the mean-median tends to be negative whereas when the cartilage has greater degree of damage, the sign of the mean-median tends to be positive. This preliminary result suggests that there could be significant relationship between these parameters which can be exploited to quantitatively differentiate cartilage that is normal and cartilage that has damage after segmentation has been performed.