Understanding the use of thermography and its ability to predict ultrasound-detected joint inflammation at the metacarpophalangeal joint in patients with rheumatoid arthritis

Objectives

To study the correlation of thermography with ultrasonography and whether thermography can help predict ultrasound-detected joint inflammation at the metacarpophalangeal joints (MCPJs) in patients with rheumatoid arthritis (RA).

Methods

Thermography recorded maximum (Tmax), average (Tavg) and minimum (Tmin) temperatures which were summed for the MCPJs of each hand, and their relationship with the summed power Doppler (PD) and grey-scale (GS) scores explored using correlation analysis and simple linear regression. The ability of the summed thermographic temperatures to predict summed PD score≥1 and summed GS score≥18 (median score) were studied using receiver operating characteristic (ROC) analysis. Intra-observer reliability (single observer) was analyzed using intra-class correlation coefficient (ICC).

Results

This cross-sectional study examined 810 joints from 81 RA patients. At both right and left MCPJs, all the summed thermographic temperatures correlated significantly (P<0.05) and had significant relationship (P<0.05) with the summed ultrasound scores (for PD and GS: correlation coefficients ranged from 0.45 to 0.52 and 0.26 to 0.29, respectively, while regression coefficients ranged from 0.094 to 0.137 and 0.058 to 0.086, respectively). At the bilateral MCPJs, the area under the ROC curves (AUCs) for the summed thermographic temperatures in predicting summed PD score≥1 and summed GS score≥18 ranged from 0.80 to 0.82 and 0.65 to 0.66, respectively. The ICC values (for 45 baseline MCPJs whose thermographic temperatures were re-segmented >2 weeks apart) were excellent (all >0.90).

Conclusions

Thermographic temperatures are reflective of ultrasound-detected joint inflammation, and appear useful in predicting PD vascularity at the MCPJs of patients with RA.