Mylène Jansen

Comparing radiographic JSW and MRI cartilage thickness using CT 241 12 Figure 1 : The 3 different imaging methods used for (in)direct cartilage quantification. The key characteristics are listed underneath each modality, and key differences between modalities are displayed in gray. For CT, both 3D and 2D joint space width measurements were used, for comparison with MRI and radiography respectively. Radiography (x-rays) Standardized weight-bearing, semi-flexed, posteroanterior (PA) radiographs were performed according to the Buckland-Wright protocol. 29,30 An aluminum step wedge was used as a reference standard to calculate the pixel size. For analysis of the radiographs, ‘knee images digital analysis’ (KIDA) software was used by 1 experienced observer, blinded to the acquisition order. The mean JSW of the most affected compartment (MAC) was calculated by averaging the tibia-femur distance at 4 locations of the MAC, which were determined automatically based on a framework of 4 lines placed manually around joint. A detailed explanation of the KIDA mathematical method has been provided in the original article. 31 MRI 3T MRIs with 3D spoiled gradient recalled imaging sequence with fat suppression (SPGR-fs) were acquired for analysis of cartilage structure using Chondrometrics Works 3.0 software. 32 Experienced observers blinded to acquisition order segmented the tibiofemoral cartilage throughout the joint, which was averaged to calculate the mean cartilage thickness of the MAC. CT Axial CT scans of the knee were performed, from which coronal reconstructions with 2 mm slice thickness were rendered. A segmentation and 3D JSW measurement method was