Table 2.
Overview of included studies
| Study | Content | Construct | Criteria | Feasibility | Reliability | Discrimination | Responsiveness |
|---|---|---|---|---|---|---|---|
| Gargani et al. [23] | 33 SSc, including 10 dcSSc, and 23 lcSSc | LUS HRCT (gold standard) PFT |
72 ScS Total BN > 10 was defined positive WS |
100% | Intra-observer and inter-observer variability respectively 5.1% and 7.4% | Total BN correlated with WS (r = 0.72; p < 0.001) and DLCO (r = – 0.6, p < 0.05); BN higher in dcSSc than in lcSSc (73 ± 66 vs 21 ± 35; p < 0.05) | N/A |
| Gutierrez et al. [25] | 28 SSc, 2 SS, 2 DM, 2 ASS, 1UCTD, and 1 MCTD | LUS HRCT (gold standard) |
50 ScS (comprehensive method) 14 ScS (simplified method); comprehensive BS was defined 0 = normal (<10 BN); 1 = mild (11–20 BN); 2 = moderate (21–50 BN), and 3 = marked (>50 BN). Simplified BS was defined 0 = normal (<5 BN); 1 = mild (6–15 BN); 2 = moderate (16–30 BN), and 3 = marked (>30 BN) WS |
Simplified method required less time than the comprehensive (8.6 ± 1.4vs 23.3 ± 4.5 minutes, p < 0.001) | κ values for inter-observer reliability of comprehensive method 0.85–0.98. κ values for inter-observer agreement of simplified method 0.77–0.89 and for intra-observer 0.85–0.89 |
BS of two methods correlated to WS (p = 0.0006), and simplified score also correlated to comprehensive method (p = 0.0001) | N/A |
| Barskova et al. [29] | 58 SSc, including 32 VEDOSS | LUS HRCT (gold standard) |
72 ScS BN ≥ 3 was found in at least two adjacent scanning sites or when total BN > 5 |
100% | Intra-observer and inter-observer variability respectively 5.1% and 7.4% | Total BN significantly higher in SSc + ILD (57 ± 53 vs 9 ± 9, p < 0.0001), and with GGO (63 ± 47 vs 33 ± 40, p < 0.05) Sensitivity and specificity respectively 100% and 55%; NPV and PPV 100% and 78% respectively |
N/A |
| Tardella et al. [28] | 26 SSc, 2 SS, 1 UCTD, 2 ASS, 2 DM, and 1 MCTD | LUS HRCT (gold standard) PFT |
50 ScS Grading as comprehensive assessment (Gutierrez et al. [25]) |
Yes | Overall agreement of inter-observer 92–97%; weighted κ value 0.85–0.98 | BS correlated with WS (p < 0.001; CC ρ = 0.875), and DLCO(p = 0.014) | N/A |
| Moazedi-Fuerst et al. [31] | 25 SSc and 40 healthy controls | LUS HRCT (only for patients) |
18 ScS Positive area for B lines = ScS with predominant B lines; positive area for PI = predominant PLT > 2.8 mm BS: 0 = no positive areas; 1 = 1–5 positive areas; 2 = > 5 positive areas. Idem for PI |
N/A | N/A | SSc + ILD had a BS of 2 in 55% and 1 in 45%; SSc – ILD had a BS of 2 in 5% and BS of 1 in 30%; SSc + ILD had a PI of 2 in 23% and 1 in 78%; SSc – ILD had a negative PI | N/A |
| Pinal Fernández et al. [38] | 21 ASS | LUS HRCT (gold standard) |
72 ScS Percentage of positive B-lines calculated (dividing the positive points by studied points) WS |
N/A | κ value of intra-observer and inter-observer 0.83 and 0.76 | BS no correlated with WS (CC = 0.135, p = 0.559); BS correlated with GGO (ρ = 0.502, p = 0.02) | N/A |
| Cogliati et al. [36] | 39 RA | LUS (standard and PS-USD) HRCT (gold standard) PFT |
72 ScS BN > 10 identified positive WS |
Yes | r-value for inter-observer variability 0.96; κ coefficient of two devices 0.78 | BS correlated with WS (r = 0.806). Sensitivity and specificity of standard LUS vs HRCT 92% and 56%, and PS-USD vs HRCT 89% and 50% | N/A |
| Moazedi-Fuerst et al. [35] | 64 RA and 40 healthy controls | LUS HRCT (gold standard, only for patients) |
18 ScS Grading as previous report (2012) |
Yes | κ value of inter-observer 0.92 | Sensitivity and specificity of LUS respectively 97.1% and 97.3%; PPV and NPV 94.3% and 98.6% respectively (p < 0.001) | N/A |
| Mohammadi et al. [26] | 70 SSc | LUS (modified TTUS) HRCT (gold standard) |
10 ScS BS: 0 = normal (≤5 BN); 1 = mild (6–15 BN); 2 = moderate (16–30 BN); 3 = severe (>30 BN) WS |
Yes | κ value of intra-observer reliability 0.838. Agreement between LUS and HRCT 0.553 (p < 0.001) |
BS correlated with WS (SCC = 0.695, p < 0.001); sensitivity and specificity respectively 73.5% and 88.23%; PPV and NPV 95.12% and 51.72% respectively | N/A |
| Gigante et al. [24] | 39 SSc, including 24 dcSSc and 15 lcSSc | LUS HRCT (gold standard) PFT NVC |
BN ≥ 3 was found in at least two adjacent scanning sites or when a total BN > 5 WS |
N/A | Intra-observer variability 3.8% | BN correlated with WS (r = 0.81, p < 0.0001), DLCO (r = – 0.63, p < 0.0001), and DSS (r = 0.8, p < 0.01) | N/A |
| Moazedi-Fuerst et al. [33] | 25 RA, 14 SSc, 6 SLE, and 40 healthy controls | LUS HRCT (gold standard, only for patients) |
18 ScS BN > 2 and PLT > 3 mm in any scanned area regarded as abnormal. Grading as previous report (2012) |
Yes | N/A | Sensitivity and specificity of LUS respectively 86.9% and 100%; PPV and NPV 100% and 88% respectively | N/A |
| Pinal-Fernandez et al. [32] | 16 SSc and 21 ASS | LUS HRCT (gold standard) PFT |
72 ScS Percentage of positive BN and PI calculated (dividing the positive points by studied points) WS and Wells scores |
N/A | N/A | PI correlated with WS both in SSc (r = 0.6, p = 0.01) and ASS (r = 0.6, p = 0.005), higher performance than BS (p = 0.01). PI also correlated with Wells score (r = 0.7, p < 0.001) and with DLCO (r = – 0.5, p = 0.05) in SSc, and high diagnostic value for detecting ILD (AUC = 0.85, 95% CI 0.64–1) and classified limited and extensive (AUC = 0.81, 95% CI 0.57–1) | N/A |
| Buda et al. [30] | 52 ILD (including 30 CTD and 16 IP) and 50 healthy controls |
LUS HRCT (gold standard) |
BN classified into three types: single (≤3 per one scan), numerous (≥4), and white lung. PI was described ragged, fragmentary, thickness (≥2 mm), and blurred WS |
N/A | N/A | Sensitivity and specificity of white lung to GGO.95% and 99%; blurred pleural line to honeycombing 59% and 82% (p < 0.005); numerous B lines correlated with blurred pleural line (p<0.001) | N/A |
| Sperandeo et al. [34] | 175 SSc | LUS HRCT (gold standard) |
BN > 3; PLT > 3.0 mm; HRCT pattern classified: no fibrosis, reticular, reticular-nodular, and honeycombing + reticular-nodular pattern | N/A | κ value of inter-observer and intra-observer 0.6–0.8 | Sensitivity and specificity of PLT (>3.0 to ≤ 5.0) to reticular pattern 80% and 99% (AUC = 0.95); PLT (>3.5) to reticular nodular and honeycombing 74.3% and 99% (AUC = 0.99); PLT (>5) to honeycombing 90.1% and 99% (AUC = 0.99) Sensitivity and specificity of BS to all kinds of fibrosis 0.94 and 0.95 |
N/A |
| Vasco et al. [37] | 13 SS | LUS HRCT (gold standard) |
8ScS BN ≥ 3 in a single ScS | N/A | κ value of intra-rater 1 | BS correlated with HRCT (r = 0.84, p < 0.001); sensitivity and specificity respectively 100% and 89% | N/A |
ASS anti-synthetase syndrome, AUC area under curve, BN B-line number, BS B-line score, CC correlation coefficient, CI confidence interval, CTD connective tissue disease, dcSSc diffuse cutaneous SSc, DLCO diffusion capacity for carbon monoxide, DM dermatomyositis, DSS disease severity scale, GGO ground glass opacity, HRCT high-resolution computed tomography, ILD interstitial lung disease, IP idiopathic pneumonia, lcSSc limited cutaneous SSc, LUS lung ultrasound, MCTD mixed connective tissue disease, N/A not applicable, NPV negative predictive value, NVC nailfold video capillaroscopy, PFT pulmonary function test, PI pleural irregularity, PLT pleural line thickening, PPV positive predictive value, PS-USD pocket size ultrasound device, RA rheumatoid arthritis, ScS scanning sites, SLE systemic lupus erythematosus, SCC Spearman’s correlation coefficient, SS Sjögren’s syndrome, SSc systemic sclerosis, SSc + ILD SSc with ILD, SSc – ILD SSc without ILD, TTUS transthoracic ultrasound, UCTD undifferentiated connective tissue disease, VEDOSS very early diagnosis of systemic sclerosis, WS Warrick score