Effect of gradient and section orientation on quantitative analysis of knee joint cartilage

The object of this study was to determine the influence of the gradient and section orientation on cartilage thickness and volume measurements in the knee joint. Eight specimens were imaged with a fat-suppressed gradient-echo sequence, applying sagittal, transverse, and coronal section orientations. Images were additionally acquired with exchanged gradient directions, and with computed tomography (CT) arthrography. After segmentation and three-dimensional (3D) reconstruction, the volume, the mean, and the maximal 3D cartilage thickness were computed. No effect of changes in the gradient orientation was found, suggesting that susceptibility-induced geometric distortion is not a relevant problem in quantitative cartilage imaging. Sagittal images produced similar data to that obtained with transverse (patella) or coronal (tibia) sections, demonstrating that all knee joint cartilages can be accurately quantified from a single sagittal data set. Whereas no significant systematic deviation between magnetic resonance imaging (MRI) and CT arthrography was recorded in the patella, there was a 10%-15% underestimation of tibial cartilage thickness in MRI.