Tandem mass tag (TMT)-based proteomic analysis was performed on plasma samples collected from 212 RA patients, of which 204 had post-treatment samples. Additionally, 56 individuals at risk of RA (pre-RA) and 99 healthy individuals were included in the study. Through the integration of bioinformatics and machine learning, the study explored protein signatures linked to pre-RA and RA, as well as predictive indicators of response to csDMARDs