A proposal for a new lung ultrasound score in rheumatoid arthritis: the reliability of lung ultrasound for rheumatoid arthritis-associated interstitial lung disease diagnosis

Abstract

Background/Objectives: The objective of this study is to investigate the intra- and inter-explorer reliability of different lung ultrasound (LUS) scores in patients with suspected rheumatoid arthritis with associated interstitial lung disease (RA-ILD). Methods: Fourteen consecutive patients with suspected RA-ILD based on the presence of respiratory symptoms, lung function tests (LFTs) or imaging alterations were recruited. The screening protocol consisted of an LFT, a chest X-ray, and HRCT. LUS examinations of different B-line and pleural line scores including 14 intercostal spaces were performed by three experienced sonographers, guaranteeing blinding. Intra- and inter-explorer reliability were calculated for all LUS scores and at the intercostal space level by weighted Cohen's kappa and Fleiss' kappa, respectively, relying on absolute differences using Stata/IC 14.2 software (R) (StataCorp, College Station, TX, USA). Results: Both global (& kgreen; = 0.73-0.82) and binary (& kgreen; = 0.80-0.90) scores of B lines showed substantial to excellent intra-explorer reliability, with slightly better results for the binary score. The inter-explorer reliability was equally excellent for the global score (& kgreen; = 0.93) and the binary score (& kgreen; = 0.90) of B lines. The intra-explorer reliability of the semiquantitative pleural score was excellent for all the sonographers (& kgreen; = 0.88-0.91), and the binary pleural score had slightly lower intra-explorer reliability (& kgreen; = 0.77-0.84). Regarding inter-explorer reliability, both semiquantitative and binary pleural scores were equally excellent (& kgreen; = 0.84). Good to excellent inter-explorer reliability was found in all the scanned areas. Conclusions: Substantial to excellent intra- and inter-explorer reliability of different feasible B-line and pleural LUS scores were found, adding evidence in favor of the potential implementation of LUS for RA-ILD diagnosis in clinical practice.