ORIGINAL ARTICLE Orthopaedics Digital scoring of haemophilic arthropathy using radiographs is feasible N. W. D. JANSEN,* K. L. VINCKEN, à A. C. A. MARIJNISSEN,* B. LUNDIN,§ L. HEIJNEN, G. ROOSENDAAL, H. PETTERSSON§ and F. P. J. G. LAFEBER* *Rheumatology & Clinical Immunology; Hematology and Van Creveld Clinic; àImage Sciences Institute, University Medical Center (UMC), Utrecht, The Netherlands;§Radiology, Lund University Hospital, Lund, Sweden; and –Rehabilitation Center De Trappenberg, Huizen, The Netherlands Summary. Radiographs are important tools to eval- uate structural changes in many joint diseases. In the case of haemophilic arthropathy (HA), the Pettersson score is widely used. The rising of digital radiography enables evaluation of these changes in a more quantitative and detailed manner, potentially improving diagnosis and follow-up. The aim of this study was to evaluate whether digital image analysis in the case of HA is feasible, using a presently available method for radiographic changes in knee osteoarthritis (OA), knee image digital analysis (KIDA). Sixty-two knee radiographs were scored according to Pettersson and with KIDA, each by two independent observers. Inter-observer variation and correlations between the two scoring methods were determined. The inter-observer variation was smaller for KIDA than for Pettersson and for KIDA not significantly different from evaluation of OA joints. Good correlations were found for the two methods where comparison of parameters was appropriate. Importantly, for each of the parameters within one point in the ordinal Pettersson score, a large window still existed in the continuous KIDA grading. Digital analysis of radiographs to quantify joint damage in HA is feasible. The use of continuous variables, as used in a digital method such as KIDA has the advantage that it enables objective and much more sensitive detection of small changes than by use of an ordinal analogue method such as the Pettersson score. Based on the present results, it would be worthwhile to adapt the KIDA method for the specific characteristics of HA and to extend the method to elbow and ankle radiographs. Keywords: digital image analysis, haemophilic arthropathy, Pettersson score, radiography Introduction Joint destruction due to recurrent haemarthroses is the main cause of morbidity in patients suffering from haemophilia [1]. Destruction of joints because of joint bleeds involves subchondral changes, enlargement of the epiphysis, changes at the joint margins, narrowing of the joint space and ultimately joint deformity [2–6]. This haemophilic arthropathy (HA) can be visualized using various imaging tech- niques, including radiography. At present, the Pet- tersson scoring system [4,7] is the most widely used method to quantify structural changes in joints of patients with haemophilia with the use of radio- graphs. This scoring system was first published in 1980 and since then has been used extensively in various types of studies. It uses radiographs of the most affected joints in haemophilia: the knees, elbows and ankles [8], which are scored for eight features of HA (Table 1). This score is on an ordinal scale (0 – max 2, for each item), and gives a maximum score for each joint of 13, yielding a maximum patient-score of 78, including all six index joints. Radiography is well established and still remains the gold standard for assessment of haemophilic arthropathy. Radiography is relatively cheap, fast, and available in almost all hospitals, even in most developing countries. There is a shift in radiography from analogue-to- digital images in radiography. The advantage is that Correspondence: Nathalie W. D. Jansen, MSc, PhD, Rheumatology & Clinical Immunology, University Medical Center, Utrecht, Room F.02.127, PO Box 85500, 3508 GA, Utrecht, The Netherlands. Tel.: 31 88 7557569; fax: 31 30 2523741; e-mail: n.w.d.jansen@umcutrecht.nl Accepted after revision 11 May 2008 Haemophilia (2008), 14, 999–1006 DOI: 10.1111/j.1365-2516.2008.01799.x Ó 2008 The Authors Journal compilation Ó 2008 Blackwell Publishing Ltd 999