RA-ILD in the Age of Treat-to-Target: Put the Probe on the Chest Save

Rheumatologists have become remarkably good at controlling joint inflammation in rheumatoid arthritis (RA). Treat-to-target strategies work. Remission is achievable. But at EULAR 2026, speakers of the session “Catching Your Breath: Unravelling RA Associated Interstitial Lung Disease (ILD)” noted it is one of the few outcomes not improving in the biologic era. Three abstracts reinforce that message and make the case for a practical, low-cost screening tool that is already in most rheumatologists’ hands.

The Burden Has Not Gone Away

The ARCTIC trial enrolled patients with newly diagnosed RA who received tight-control, treat-to-target therapy from disease onset in line with current EULAR recommendations. At the ten-year mark, Mangseth and colleagues examined HRCT of 170 patients reviewed centrally by expert thoracic radiologists with multidisciplinary team confirmation (POS1270). Despite 82% of patients being in DAS remission at the time of imaging, 9.4% had confirmed RA-ILD and an additional 19.4% showed other non-specific interstitial findings. Only 71% had completely clear lungs. In regression analyses, male sex and older age were associated with increased RA-ILD risk, while seropositivity was associated with lower risk, a counterintuitive finding that challenges the assumption that ACPA-positive patients carry the highest burden.

HRCT remains the diagnostic gold standard, but its radiation exposure, cost, and risk of incidental findings create real barriers to routine use as a screening tool. The field needs a practical alternative, particularly given that risk factor-based criteria flag the vast majority of RA patients for imaging under current ACR/CHEST and ERS/EULAR guidelines.

Thoracic Ultrasound as a First-Pass Screen

Recalde-Reyes and colleagues from Universidad Nacional de Colombia evaluated 147 RA patients using a standardized 14-point lung ultrasound protocol performed by trained rheumatologists, with HRCT as the reference standard (OP0345). Among asymptomatic patients, lung ultrasound achieved 95.4% sensitivity and 77.7% specificity for RA-ILD detection, with a negative likelihood ratio of 0.06. Crucially, clinical risk scores added almost nothing: the Spanish consensus score produced a positive likelihood ratio of just 0.97, barely better than chance. Combining lung ultrasound with risk scores pushed sensitivity to 100%. Lung ultrasound outperforms clinical risk scores for detecting RA-ILD, even in patients without respiratory symptoms. The single-center design limits generalizability, but the effect sizes are large enough to be signal, not noise.

Those findings are reinforced by a multicenter pooled analysis across Argentina, Denmark, and Belgium (POS1265). Otaola and colleagues analyzed 272 patients using a harmonized 14-zone thoracic ultrasound protocol evaluating total B-line score and pleural line irregularities against HRCT. Prevalence of RA-ILD in the cohort was 23.5%. A B-line score greater than 5 provided the best diagnostic balance: 85.9% sensitivity, 73.1% specificity, and a negative predictive value of 94.2%. The AUC was 0.86. Adding pleural line irregularities raised sensitivity to 96% and NPV to 95%, at the cost of lower specificity. In practical terms, a negative thoracic ultrasound makes RA-ILD very unlikely.

A Stepwise Screening Pathway

Read together, these data support a tiered approach: screen with thoracic ultrasound, confirm positives with HRCT, and reserve CT for patients who truly need it. Rheumatologists who already use musculoskeletal ultrasound could extend that practice to a standardized lung protocol, enabling a preliminary screening assessment without waiting for a pulmonology referral. The point-of-care advantage is particularly meaningful for older male patients, the subgroup most consistently identified as high-risk, in whom a brief thoracic ultrasound integrated into routine monitoring visits could enable earlier identification before DLCO decline and overt dyspnea develop.

Operator dependency remains a real limitation, and standardized training curricula will be essential before broad adoption is feasible. AI-assisted interpretation may be possible with the kind of harmonized 14-zone protocols validated here, to address some of the variability and accessibility of thoracic ultrasound as a screening tool.

What Comes Next

Among the possible options for RA-ILD screening, including low-dose CT, lung MRI, circulating biomarkers such as MMP-7 and KL-6, and MUC5B genetic profiling, thoracic ultrasound stands out as a practical and immediately actionable option. It is non-ionizing, low cost, already in many rheumatologists’ hands, and now validated across multiple international cohorts. Prospective longitudinal data linking early thoracic ultrasound-based detection to patient outcomes are the critical next step. The tools to find RA-ILD earlier are arriving. But as Prof. Anna-Maria Hoffmann-Vold noted in the “Catching Your Breath” session, “screening is only as good as the early intervention we have available” and the harder question of what to do once we find it remains very much open.