Impact of quantitative radiological features of interstitial lung disease on immunomodulatory treatment response in three autoimmune interstitial lung disease cohorts

Background The defining radiological features of autoimmune interstitial lung disease (ILD) are ground glass opacification (GGO) and fibrosis. The associations between these features and physiological response to immunomodulation remain unclear. Methods This study leveraged three autoimmune ILD cohorts: two with systemic sclerosis (SSc) and one with rheumatoid arthritis (RA) which were selected for inherent differences in fibrotic extents/patterns. Linear regression models examined associations between baseline quantitative GGO, fibrosis, their ratio and forced vital capacity (FVC)%-predicted changes after 12 months of immunomodulatory therapy. Results Patients with SSc-ILD (N=262) exhibited a higher GGO-to-fibrosis ratio compared with patients with RA-ILD (N=130) (mean ratio 3.0 vs 0.25). Increased GGO-to-fibrosis was not associated with improved FVC%-predicted in any cohort. Conversely, in patients with SSc-ILD treated with cyclophosphamide (CYC), increased fibrosis (estimate 0.17 (95% CI 0.003, 0.33); p=0.04) and increased GGO (estimate 0.15 (95% CI 0.004, 0.30); p=0.044) were both significantly associated with FVC% improvement. Given the negative direction of the estimate for GGO-to-fibrosis ratio (estimate −0.33 (95% CI −0.61, –0.06); p=0.016), CYC was associated with greater FVC% improvement in patients with a higher degree of fibrosis relative to GGO. No significant correlation was seen in patients with SSc-ILD treated with mycophenolate (N=56) or in patients with RA-ILD treated with immunomodulation (N=130). Discussion Increased quantitative GGO relative to fibrosis was not significantly associated with improved response to immunomodulation in patients with RA-ILD and SSc-ILD. However, increased quantitative fibrosis and GGO extent were associated with improved response to CYC in SSc-ILD. More research is needed to understand how to use radiological features to guide treatment selection in ILD. Data are available upon reasonable request.