SAT0572 CONSTRUCTIVE VALIDITY OF MUSKULOSKELETAL ULTRASOUND MEASUREMENT OF CARTILAGE THICKNESS IN PATIENTS WITH KNEE OSTEOARTHRITIS

Abstract

Background: Cartilage thickness is one important measure in describing both OA development and progression. Based on current knowledge, conventional radiography (CR) and magnetic resonance imaging (MRI) have not been demonstrated to be superior over one another. Because of disadvantages of MRI and CR neither can be use in routine daily clinical practice for follow up of OA patients. Diagnostic ultrasound assessment (US) of cartilage thickness offers an alternative measure as a clinically available and more cost-effective source of knee articular cartilage imaging. Objectives: Our objective was to determine the relationship between US and CR measures of femoral cartilage thickness in patients with knee osteoarthritis because systematic feature- and site-specific cross-comparison between this two methods is still missing in the current literature. Methods: 120 patients with knee osteoarthritis (240 knees) are recruited for this study. The joint space width (JSW) and Kellgren and Lawrence (K&L) grade were measured using weight-bearing anteroposterior 30° knee semi-flexion knee radiography (with inclusion criteria K&L grade 1-4). Femoral cartilage thickness was measured three times in supine position and with a suprapatellar transverse scan with the knee in maximal flexion at the lateral condyle (LC), medial condyle (MC) and intercondylar notch (IN) by one rheumatologist and arithmetic mean is taken. Pain andfunctionality are measured with VAS pain scale, Womac, Lysholm and SF 36 score. The agreement between two methods was evaluated with Bland-Altman analysis. Results: We found a statistically significant low level of rank correlation between CR and US measurements of mean cartilage thickness; ρ (rho) values between modalities were low (0.263 and 0.273 depending on side (right/left), p=0.005 and p=0.007 respectively). In Bland - Altman analysis, US measurement showed bad agreement with CR. Presence or absence of US features of OA (effusion, synovial hypertrophy, osteophytes and popliteal cysts) didn't influence on cartilage thickness assessed by US (p>0.05). For US assessment, we found correlation only between cartilage thickness and VAS pain scale (ρ (rho) -0.281, p=0.004). We didn't found any statistically significant correlation between CR thickness measures and pain/functionality/HRQoL scores (p>0.05). Conclusion: These results suggest that ultrasound may be a useful clinical tool to assess relative cartilage thickness. However, the absolute validity of the ultrasound measure is called into question due to the larger CR-based thickness measures and low level of agreement according to Bland-Altman analysis. The use of ultrasound as a complementary imaging tool along with CR may enable more accurate and cost-effective detection, prognosis and follow-up of knee osteoarthritis in routine clinical practice.