Background The clinical application of quantitative measurement of separate radiographic parameters of knee osteoarthritis (OA) might be hampered by a lack of reproducible joint positioning during acquisition of the radiographs.
Objectives To evaluate which systematic variations in positioning of the knee towards the X-ray detector have an effect on the measurement of separate parameters by Knee Images Digital Analysis (KIDA).
Methods Compared to the semiflexed Buckland-Wright position 5 components of knee position during acquisition (beam height, lower and upper leg extension, internal rotation, and lateral shift) were systematically varied within a clinically relevant range, using three cadaver legs. KIDA was performed to measure joint space width (JSW), varus angle, eminence height, osteophyte area, and bone density in the knee. Using linear regression analyses the influence of the systematic variations on the KIDA parameters was evaluated. Significant changes were validated in vivo in healthy individuals, and changes were compared with differences after 2 years in a radiographic progression cohort of early OA (K&L grade < II at baseline, and K&L ≥II five years later; n=310 knees).
Results Systematic variation in upper leg extension decreased lateral JSW by 0.20mm per 5 degrees extension, which was only slightly smaller than the 0.27mm difference after 2 years in the knees with early OA that progressed. Variation in lower leg extension increased minimum JSW by 0.07mm per cm forward shift of the lower leg, which is clinically relevant compared to the 0.11mm difference in knees with OA progression. This variation also influenced lateral eminence height; the decrease of 0.23mm per cm was only slightly smaller than the 0.27mm in knees with OA progression. These influences were all confirmed in vivo. Additionally, an influence on the different KIDA parameters was found that was confirmed in vivo, but that was smaller than the differences in knees with OA progression. E.g. by varying upper leg extension the varus angle decreased (cadaver: 0.25, in vivo: 0.08 degrees per 5 degrees). Varying lower leg extension influenced minimum (0.07mm, in vivo 0.17mm per 5 degrees), medial (0.07mm, in vivo 0.24mm), and lateral JSW (-0.18mm, in vivo -0.03mm), and varus angle (-0.32 degrees, in vivo -0.31 degrees). Lower leg extension also influenced medial tibia osteophyte area (0.49mm2, in vivo 0.22mm2), and medial eminence height (-0.17 vs. -0.04mm in vivo). Although smaller than the differences due to OA progression the bone density was influenced by variation in all 5 position components.
Conclusions Variations in knee positioning, which easily occur during image acquisition in trials and clinical practice despite standardization, significantly influence the measurement of most radiographic OA features. Although the digital parameters are sufficiently robust, their surplus value over qualitative grading will pay off only when standardization during image acquisition is improved. Since radiography remains cheap and easily accessible, it is considered of value to further improve standardization of acquisition.
Disclosure of Interest None Declared