Article Text

AB1168 Clinical disease activity scores of knee joints in newly diagnosed JIA predicts reduction in bone pixel value detected by computed radiography
  1. L. Pascoli1,
  2. A. Mc Cann2,
  3. M. Stevenson3,
  4. C. Mc Allister4,
  5. M. Rooney5
  1. 1Rheumatology, Azienda Ospedaliera San Camillo-Forlanini, Roma, Italy
  2. 2Regional Medical Physics Service
  3. 3Statistics, Queen’s University
  4. 4Paediatric Rheumatology, Musgrave Parck Hospital
  5. 5Paediatric Rheumatology, Queen’s University, Belfast, United Kingdom


Objectives To evaluate the relationship between Computed Radiography (CR) Pixel values and disease activity scores obtained for the knees of newly diagnosed Juvenile Idiopathic Arthritis (JIA) children.

Methods Children with newly diagnosed untreated JIA were recruited. Both knees were scored for disease activity on a score of 0-3. Weight-bearing antero-posterior (AP) images of both knees were captured via CR. A hydroxyapatite phantom was fixed to an unobscured region of the phosphor cartridge during all exposures. Pixel values in the phantom region were used to determine the brightness and contrast settings of the CR system allowing conversion of the whole image to values of absolute attenuation. Scattered radiation, whose random nature interferes with measurement of tissue content, was estimated based on Monte Carlo N-particle (MCNP) simulations. Measurements of bone and soft tissue content were taken at 9 repeatable identifiable regions in AP X-rays of both knee joints. The effect of overlying soft tissue in the bone regions was compensated by extracting the residual values of bone versus soft tissue plots over the full data set. Knee joints for all patients were independently assigned clinical scores (none, mild, moderate, severe). Association of clinical score categories with soft tissue and bone residual parameters was characterized by ordinal regression with the individual subject incorporated as a blocking variable.

Results Forty children were recruited and a total of 80 knees’ digital images were obtained. Results showed that a higher clinical score correlates with reduced “bone density” (p=0.003) and with greater ST content (p<0.0001). Moreover, the “bone loss” occurs only in the epiphyseal region compared to the diaphyseal region. Furthermore, we found good agreement between observed and predicted clinical scores on the basis of the “bone density” at the epiphyseal region (Kendall’s tau 0.61).

Conclusions We have demonstrated that a higher clinical score correlates with a reduced “bone density” in the periarticular epiphyseal regions, which is caused by the inflammation process. Thus, following appropriate image acquisition and analysis, CR images can be used to measure periarticular bone loss in JIA. The results obtained from standard X-rays could be of substantial value to the clinician using serial X-rays to determine joint damage over time. We are currently evaluating the CR images obtained over the 2 years of our prospective study and comparing these with disease outcome during the same time period.

Disclosure of Interest None Declared

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.