Comhaire FH, Schoonjans FA. Royal Cynological Society Saint Hubert, Albert Giraudlaan, 98, B 1030 Brussels, Belgium Ghent University Hospital, Department of Endocrinology, De Pintelaan, 185, B 9000 Gent, Belgium. JSAP 2011 Oct 52 10 536-42 Objectives: To assess the reproducibility of the Fédération Cynologique Internationale (FCI) classification over time, and between assessors and to evaluate the benefit of computer-assisted measurement of the Norberg angle and of its percentile ranking by breed for the diagnosis of canine hip dysplasia, and for the selection of couples for breeding. Methods: During the time period of 2006 to 2010, 5094 hip radiographs were evaluated according to the FCI rules, and 4400 have been submitted for statistical analysis. A system of computer-assisted image analysis (Digimizer(®) , MedCalc Software Ltd, Mariakerke, Belgium) was used to measure the Norberg angle. The Norberg angle value of individual dogs was expressed as percentile rank by breed. Results: The agreement between individual assessors was highly significant (P<0•001), but there were important variations over time of the ratio of classes A or B. The Norberg angle and the percentile rank accurately discriminated between dogs with or without canine hip dysplasia, with positive and negative likelihood ratios of 6•31 and 0•21, respectively, for the Norberg angle at criterion value of 102•2°, and 4•21 and 0•18 for the percentile rank at criterion value of the 25th percentile. Clinical Significance: The percentile rank of the Norberg angle may be a valuable tool for breeding selection.

• Arthritis
• Hip
• Imaging
• Osteoarthritis