MRI based knee cartilage assessment

Abstract

Osteoarthritis is one of the leading causes of pain and disabilityworldwide and a major health problem in developed countries due to thegradually aging population. Though the symptoms are easily recognizedand described by a patient, it is difficult to assess the level ofdamage or loss of articular cartilage quantitatively. We present a novelmethod for fully automated knee cartilage thickness measurement andsubsequent assessment of the knee joint. First, the point correspondencebetween a pre-segmented training bone model is obtained with use ofShape Context based non-rigid surface registration. Then, a singleActive Shape Model (ASM) is used to segment both Femur and Tibia bone.The surfaces obtained are processed to extract the Bone-CartilageInterface (BCI) points, where the proper segmentation of cartilagebegins. For this purpose, the cartilage ASM is trained with cartilageedge positions expressed in 1D coordinates at the normals in the BCIpoints. The whole cartilage model is then constructed from thesegmentations obtained in the previous step. An absolute thickness ofthe segmented cartilage is measured and compared to the mean of alltraining datasets, giving as a result the relative thickness value. Theresulting cartilage structure is visualized and related to the segmentedbone. In this way the condition of the cartilage is assessed over thesurface. The quality of bone and cartilage segmentation is validated andthe Dice's coefficients 0.92 and 0.86 for Femur and Tibia bones and 0.45and 0.34 for respective cartilages are obtained. The clinical diagnosticrelevance of the obtained thickness mapping is being evaluatedretrospectively. We hope to validate it prospectively for prediction ofclinical outcome the methods require improvements in accuracy androbustness.