Agreement and reliability of thoracic ultrasound findings associated with interstitial lung disease in rheumatoid arthritis: a sub-study of the AURORA cohort

In thoracic ultrasound (TUS) the B-line artefact and irregular visceral pleura have been associated with the presence of interstitial lung disease (ILD). The aim of this study was to assess inter- and intraobserver reliability of TUS findings indicating ILD across different competency levels. In this reliability study derived from a primary diagnostic accuracy cohort, 80 patients with rheumatoid arthritis (RA) underwent TUS assessment. A total of 1,120 clips were independently assessed for B-lines and visceral pleural abnormalities by three blinded observers with different competency levels; European Federation for Ultrasound in Medicine and Biology (EFSUMB) levels 1 and 3, as well as point-of-care ultrasound competency level. One observer (EFSUMB level 1) repeated assessments to evaluate intraobserver agreement. Primary outcomes were percentage agreement and Cohen’s kappa (κ) with 95% confidence intervals (CI). Overall TUS positivity showed good to excellent agreement (80–91%) with κ 0.61–0.82; 95% CIs were 69–88 and 82–96 for agreement, and 0.44–0.78 and 0.70–0.95 for κ. Agreement for ≥ 10 B-lines was 84% to 94% (κ 0.68–0.87), with 95% CIs of 73–91 and 85–98 for agreement, and 0.53–0.84 and 0.76–0.98 for κ. Agreement for bilateral visceral pleural abnormalities was 65% to 83% agreement (κ 0.35–0.62), with 95% CIs of 53–75 and 72–90 for agreement, and 0.15–0.54 and 0.43–0.80 for κ. Among observers who have received structured TUS training and competency assessment, TUS signs indicating ILD are reproducible. Lower agreement for pleural abnormalities suggests that greater emphasis on pleural assessment may be warranted in TUS training. The original study is registered at clinicaltrials.gov (NCT05396469).