Mylène Jansen

256 Chapter 13 removed under anesthesia, after which patients were discharged the same day, without further imposed rehabilitation protocol. The study was approved by the medical ethical review committee of the University Medical Center Utrecht (04/086). All patients gave written informed consent. As the current study was initiated long ago, patients were not included in the design of or recruitment to the study. However, in the past years a patient council was established and multiple meetings with KJD patients have been held, with the purpose of directly involving patients in research and gathering their input on the treatment and related research. Patients from all our OA research, including the current study, receive newsletters with updates on study results. MRI analyses 1.5T MRI scans including a coronal 3D spoiled gradient recalled echo sequence with fat suppression (SPGR-fs) were acquired shortly before and at 1, 2, 5, 7, and 10 years after surgical treatment. A slice thickness of 1.5 mm, repetition time of 20 ms, echo time of 9 ms, flip angle of 15 degrees, acquisition matrix of 512x512 pixels, and pixel size of 0.31x0.31 mm were used. Images were imported into Stradview v6.0 (University of Cambridge Department of Engineering, Cambridge, UK, in-house developed software freely available at https://mi.eng. cam.ac.uk/Main/StradView) , which was used for semi-automatic cartilage segmentation. Initial contours were drawn manually for the tibia and femur every 5 slices, from which a 3D isosurface was generated for the 2 bones separately. The inner and outer cartilage surfaces were measured automatically in every slice and checked manually. Data sampled along a vector at the normal to each vertex of the surface on the cartilage patches was used to calculate the distance between the inner and outer surface and with that obtain the cartilage thickness at each vertex via model-based deconvolution. This process was performed for every scan for patches of the femur, medial and lateral tibia separately, and has previously been described in more detail. 17 The outer surface of all obtainedpatches were registered to representative canonical surfaces using an initial similarity transformation and subsequent thin-plate spline registration, performed in wxRegSurf v18 (University of Cambridge Department of Engineering, Cambridge, UK, in-house developed software freely available at http://mi.eng.cam.ac.uk/~ahg/wxRegSurf/) to allow comparing patches from multiple scans. Initial analyses focused on the whole joint (patches). To analyze the average cartilage thickness on both sides of the joint separately, relevant medial and lateral weight-bearing parts of the femur were selected (cut out) from the canonical surface (and thus applied identically in all patients and time points) in wxRegSurf (Supplementary Figure S1). An average cartilage thickness for both the femur and tibia on both sides of the joint could be generated by averaging the thickness values of all vertices in the 4 parts separately.