Abstract
Background: Interstitial lung diseases (ILDs), including idiopathic pulmonary fibrosis (IPF) and connective tissue disease-associated ILD (CTD-ILD), share similar features that complicate diagnosis. Tumor markers are often elevated in ILD, yet their diagnostic utility remains unclear. Methods: This retrospective study included ILD patients hospitalized between 2018 and 2025. Serum levels of alpha-fetoprotein, carcinoembryonic antigen (CEA), carbohydrate antigen (CA) 199, CA125, CA153, neuron-specific enolase (NSE), and cytokeratin 19 fragment (CYFRA 21-1) were analyzed. Arterial blood gases and erythrocyte sedimentation rates (ESRs) were also collected. Statistical analyses involved the Kruskal–Wallis test, Dunn’s post hoc test, Spearman’s correlation, logistic regression, and receiver operating characteristic (ROC) curve analysis. Results: CEA, CA199, and CA125 levels varied significantly among ILD subtypes (all p < 0.05). NSE differed among CTD-ILD subgroups (p = 0.0409). In IPF, CEA and NSE correlated inversely with PaO? (r = -0.1556, p = 0.0380; r = -0.2205, p = 0.0031). In CTD-ILD, NSE correlated negatively with PaCO? (r = -0.1811, p = 0.016), and CYFRA 21-1 with PaO? (r = -0.1999, p = 0.0078). A diagnostic model incorporating CEA, CA199, sex, age, smoking, PaO?, and ESR differentiated IPF from CTD-ILD with an AUC of 0.833 (95% CI: 0.790–0.876), showing 73.6% sensitivity and 82.4% specificity at a cutoff of 0.569, outperforming single markers. Conclusion: CEA, CA199, and CA125 aid in distinguishing ILD subtypes, while CEA, NSE, and CYFRA 21-1 correlate with impaired gas exchange. The combined clinical and biomarker model demonstrated superior performance in discriminating IPF from CTD-ILD, highlighting its clinical potential.

This publication has 23 references indexed in Scilit: