From:  Advances in the validation of lung ultrasound for the detection of interstitial lung disease associated with rheumatoid arthritis

 Cross-sectional and retrospective studies on the accuracy of LUS.

Author, year of publicationStudy characteristicsLUS protocolResults
Otaola et al., 2025 [41]Multicenter (5 centers), cross-sectional study including RA without respiratory symptoms (n = 203)Convex probe and linear probe
14 ICS; BL numbers and PI
ILD by LUS: BL score ≥ 5

  • LUS for ILD diagnosis: Se 83%, NPV 93.1%

  • DLCO for ILD diagnosis: Se 52%, NPV 78.6%

Zheng et al., 2025 [39]Monocentric, retrospective study including RA, regardless of respiratory symptoms (n = 120)Cardiac sector probe (2.5–3.5 MHz)
50 ICS; BL score = sum of BL
Sonographic ILD: BL > 10

  • Concordance rate between LUS and HRCT 83.33%

  • LUS accuracy: Se 86.84%, NPV 77.27%

  • Optimal cutoff: 12 BL (AUC = 0.89, p < 0.001)

  • BL scores were significantly higher in patients with ILD vs. non-ILD, and in the UIP vs. NSIP pattern

Vermant et al., 2025 [38]Monocentric, cross-sectional study including RA, regardless of respiratory symptoms (n = 116)Curved array probe (3.5 MHz)
72 ICS
Number of BL

  • Number of BL was the most strongly associated with the clinical-radiological score#

  • BL cutoff = 5: outperformed mMRC, dyspnoea, or cough

  • LUS accuracy: Se 90–100%, NPV 0.96–1

Schneeberger et al., 2025 [37]Monocentric, cross-sectional study including RA, regardless of symptoms (n = 107)Convex probe (1–8 MHz)
14 ICS
Total BL score$

  • LUS cutoff ≥ 5 BL: AUC 0.86, Se 87.1%, Sp 74.3%

  • Clinical scoreϮ cutoff ≥ 5.5: AUC 0.80, Se 75%, Sp 71%

  • Clinical score including LUS (≥ 5 BL) cutoff ≥ 7.5: AUC 0.88, Se 84.4%, Sp 75%

Reichenberger et al., 2024 [36]Monocentric, cross-sectional study including asymptomatic RF/ACPA-positive RA (n = 67)Convex probe (3.5 MHz)
14 ICS; PI and BL > 5
Suspected ILD: LUS and PFT

  • ILD suspicion by LUS and PFT: Se 60%, NPV 71%

  • ILD suspicion by LUS: Se 71%, NPV 75%

  • ILD suspicion by PFT: AUC 0.26

Tanten Zabaleta et al., 2024 [35]Monocentric, cross-sectional study including consecutive RA patients, regardless of symptoms and previous ILD (n = 104)Multifrequency linear probe (8–18 MHz) and convex probe (4–8 MHz)
14 ICS
Total number of BL and PI

  • Patients with ILD had more BL and PI (p < 0.001) than patients without ILD

  • Best cutoffs for significant ILD were 8 BL and 7 PI

  • BL ≥ 8: Se 80.9%, NPV 94.3%: PI ≥ 7: Se 80.9%, NPV 94%

Santos-Moreno et al., 2024 [34]Monocentric, cross-sectional study including RA patients with clinical ILD suspicion (n = 192)Convex probe (2–5 MHz)
72 ICS
BL and PI, semiquantitative and binary scores

  • LUS for ILD diagnosis: Se 98.3%, NPV 84.6%

  • BL cutoffs for ILD: 11.5 BL

  • Chest X-ray for ILD diagnosis: Se 29.9%, NPV 46.4%

Otaola et al., 2024 [33]Cross-sectional study including RA patients, regardless of respiratory symptoms (n = 106)Convex probe (1–8 MHz)
14 ICS
Total number of BL
LUS positive if BL ≥ 5

  • LUS for ILD diagnosis: Se 90.6%, NPV 94.7%

  • LUS accuracy was superior to PFT, chest auscultation, and chest X-ray

Di Carlo et al., 2022 [32]Monocentric, cross-sectional study including RA patients with suspicion of ILD (n = 72)Linear probe (4–13 MHz)
14 ICS
Total number of BL

  • Good positive correlation between the number of BL and the percentage of chest HRCT fibrosis

  • BL cutoff of 9 to identify significant ILD

  • LUS accuracy: Se 70%, Sp 97.6%

Cogliati et al., 2014 [31]Monocentric, cross-sectional study including RA patients with suspicion of or known ILD (n = 39)Experienced and short-trained examiners, with convex and phased array probes
Standard and pocket-sized machines, 72 ICS and 8 ICS

  • BL score significantly correlated with HRCT Warrick score (r = 0.806)

  • Kappa value: 0.78

Return to article view

#: Clinical-radiological score: included medical record data, radiological assessment, PFT, and HRCT into a 4-point score: 0 = normal, 1 = non-specific abnormalities, 2 = subclinical interstitial lung changes, and 3 = advanced ILD; $: total B-line score (TBLS): counting the number of BL if < 5, assigning a value of 5 if BL ≥ 5, and a value of 10 if the presence of white lung. LUS was considered positive if TBLS > 5; Ϯ: clinical score of 5 variables: male sex (3 points), crackles (3 points), age ≥ 60 years (2 points), positive RF (2 points), and ACPA positive antibodies (1 point); ACPA: anti-cyclic citrullinated peptide antibodies; AUC: area under the curve; BL: B lines; DLCO: carbon monoxide diffusion capacity; HRCT: high-resolution computed tomography; ICS: intercostal spaces; ILD: interstitial lung disease; LUS: lung ultrasound; mMRC: modified Medical Research Council; NPV: negative predictive value; NSIP: non-specific interstitial pneumonia; PFT: pulmonary function tests; PI: pleural ultrasound irregularities; RA: rheumatoid arthritis; RF: rheumatoid factor; Se: sensitivity; Sp: specificity; UIP: usual interstitial pneumonia.