The comparisons between thermography and ultrasonography with physical examination for wrist joint assessment in juvenile idiopathic arthritis

Abstract

© 2017 Institute of Physics and Engineering in Medicine. Objective. This study aimed to assess infrared thermography (IRT) and ultrasonography (US) for detecting wrist arthritis in juvenile idiopathic arthritis (JIA) patients. Although IRT could help us in detecting joint inflammation, IRT studies in JIA patients with wrist arthritis are still limited. Currently, no validated US criteria exist for detecting arthritis, and the most useful parameters between gray-scale ultrasound (GSUS) or power Doppler ultrasound (PDUS) remain unclear. Approach. Forty-six JIA patients were included in this study. Detecting wrist arthritis at varying degrees using IRT and US were compared with physical examination. Main results. Sixteen patients had previous wrist arthritis that is currently inactive and 30 still had wrist arthritis. The median ages (IQR) were 7.7 (4.3) and 10.2 (4.8) years, respectively. Fifteen healthy participants were included, with a median age (IQR) of 9.2 (2.0) years. Using IRT, mean temperature (T mean) and maximum temperature (T max) at skin surface in the region of interest (ROI) in the arthritis group were higher than in the inactive group and the healthy controls with p < 0.05. When patients with arthritis were subgroup analyzed by disease severity based on physical examination, the moderate to severe arthritis had T mean and T max higher than the mild arthritis group with statistical significance. The heat distribution index (HDI), two standard deviations of all pixel temperature values in the ROI, in the moderate to severe arthritis group was higher than in the healthy controls (p = 0.027). The receiver operating characteristic analysis in arthritis detection revealed diagnostic sensitivity of 85.7% and 71.4% and specificity of 80.0% and 93.3% at cut-off points of T mean 31.0 °C and T max 32.3 °C, respectively. For US, GSUS and PDUS are useful in detecting arthritis, providing high sensitivity (83.3%) and specificity (81.3%). Significance. Our study demonstrated that both IRT and US were applicable tools for detecting wrist arthritis.