Background/Purpose: Quantitative computed tomography (QCT) methods using machine learning have been developed to automatically quantify parenchymal lung features on chest CT imaging. QCT can detect subtle or early abnormalities, and abnormal QCT findings have been associated with symptoms, progressive lung disease, and mortality in general population studies. Based on prior studies linking RA disease activity to interstitial lung disease (ILD) in patients with established RA, we hypothesized that higher RA disease activity would be associated with increased interstitial changes on QCT in early RA.

Methods: We performed a cross-sectional study of participants in SAIL-RA (Study of Inflammatory Arthritis and ILD in Early RA), an ongoing prospective, multicenter longitudinal study of patients recently diagnosed with RA at five U.S. sites between 2017 and 2025. All participants had early RA (diagnosed < 2 years prior to enrollment) and were assessed with physical examination, chest high-resolution CT imaging, pulmonary function tests (PFTs), and peripheral blood testing. QCT was performed using a k-nearest neighbor machine learning classifier that categorizes regions of interest into normal lung, interstitial changes, or emphysema based on the local tissue density and distance from the pleural surface. Interstitial changes on QCT may represent features of ILD or other abnormalities such as scarring after infection. We examined the correlation of QCT interstitial changes with PFTs, RA disease activity, and erythrocyte sedimentation rate (ESR) using Spearman correlation coefficients. We tested the association between baseline age, sex, smoking status, RF, anti-CCP, and RA disease activity with QCT interstitial features using multivariable linear regression.

Results: We performed chest QCT analysis on 200 participants with early RA (median RA duration 0.72 years, mean age 55.7 years, 75% female, 85% seropositive) (Table 1). On QCT, early RA patients had mean normal lung percentage of 88.0% (SD 7.8). The mean percentage of interstitial changes and emphysema was 5.6% (SD 6.8) and 1.2% (SD 1.7), respectively. Interstitial changes on QCT were correlated with higher DAS28-ESR (rho=0.26, p=0.0003) and ESR (rho=0.30, p< 0.0001), as well as lower FVC% (rho=-0.23, p=0.002) and DLCO% (rho=0.42, p< 0.0001) (Figure 1). In a multivariable linear regression model, older age (b=+0.14 per year, 95%CI 0.07-0.21, p< 0.0001) and DAS28-ESR (b=0.64 per point, 95%CI 0.01-1.26, p=0.046) were independently associated with percentage of interstitial changes (Table 2). The b from the linear regression model can be interpreted as a percentage increase in interstitial changes per unit increase in the predictor (increase of 0.70% per point of DAS28-ESR).

Conclusion: In this prospective, multicenter study of 200 patients with early RA analyzed using machine learning QCT methods, older age and higher RA disease activity were associated with more QCT interstitial changes. QCT interstitial changes were correlated with lower FVC and DLCO. These findings emphasize the role of inflammation in RA-ILD pathogenesis early in the RA disease course and demonstrate the potential of QCT for RA-ILD screening and monitoring.

« Back to ACR Convergence 2025

ACR Meeting Abstracts - https://acrabstracts.org/abstract/disease-activity-inflammation-and-older-age-are-associated-with-quantitative-interstitial-abnormalities-in-early-ra-results-from-a-multicenter-prospective-cohort-study/