Fig. 1

Evaluation of specificity of the 13 candidate biomarkers for detecting ILD. (A) Overview of the study design. (B) Correlation matrix analysis of the 13 candidate biomarkers. The values in the heatmap represent Spearman’s correlation coefficients. (C, D) Examination of ILD specificity of 13 candidate biomarkers and excerpts on pro SP-B, KL-6, and CCL17. The values on the heatmap represent fold change compared with that in HCs. The serum levels of the candidate biomarkers were compared using analysis of variance (ANOVA), and the Dunnett’s method was applied to adjust the P values obtained via ANOVA. *, adjusted P < 0.05. Numbers of samples: HC (n = 30), BA (n = 30), COPD (n = 28), NTM (n = 15), bacterial pneumonia (n = 6), IPF (n = 86), non-IPF-ILD (n = 57). BA, bronchial asthma; CCL, C-C motif chemokine ligand; COPD, chronic obstructive pulmonary disease; HC, healthy control; ILD, interstitial lung disease; IPF, idiopathic pulmonary fibrosis; KL-6, Krebs von den Lungen-6; NTM, nontuberculous mycobacterial infection; SP-B, surfactant protein-B