Multi-examiner reliability of automated radio frequency-based ultrasound measurements of common carotid intima–media thickness in rheumatoid arthritis

Abstract

To assess the reliability of the automated radio frequency (RF)-based US measurement of carotid intima-media thickness (IMT) performed by rheumatologists and to evaluate the variability between this method and the conventional B-mode US measurement of carotid IMT in RA patients. Twelve rheumatologists measured in two blinded rounds the IMT of both common carotid arteries (CCAs) of seven RA patients with an automated RF-based method. At each round, a cardiologist measured both CCA-IMTs of the patients using an automated B-mode method. Inter-observer reliability for RF-based IMT measurements was evaluated by the intra-class correlation coefficient (ICC). Intra-observer reliability for RF-based IMT measurements was assessed using the root mean square coefficient of variation (RMS-CV), Bland-Altman method and ICC. Agreement between the two US methods was evaluated by the Bland-Altman method, ICC and RMS-CV. Inter-observer ICCs for the RF-based CCA-IMT measurements were 0.85 (95% CI 0.69, 0.94) for the first round, and 0.77 (95% CI 0.55, 0.91) for the second round. RMS-CVs for the RF-based CCA-IMT measurements varied from 5.6 to 11.7%. The mean intra-observer ICC for the RF-based CCA-IMT measurements was 0.61 (95% CI 0.46, 0.71). In the Bland-Altman analysis for agreement between RF-based and B-mode CCA-IMT measurements, the mean difference varied from -0.6 to -19.7 μm. Inter-method ICCs varied from 0.57 to 0.83 for 11 rheumatologists. Inter-method RMS-CVs varied from 11.3 to 13.7%. Our results suggest that automated RF-based CCA-IMT measurement performed by rheumatologists can be a reliable method for assessing cardiovascular risk in RA patients.