Constructing an extracellular matrix-related prognostic model for idiopathic pulmonary fibrosis based on machine learning.

backgroundIdiopathic pulmonary fibrosis (IPF) is a chronic and progressive interstitial lung disease. Multiple research has revealed that the extracellular matrix (ECM) may be associated with the development and prognosis of IPF, however, the underlying mechanisms remain incompletely understood.

methodsWe included GSE70866 dataset from the GEO database and established an ECM-related prognostic model utilizing LASSO, Random forest and Support vector machines algorithms. To compare immune cell infiltration levels between the high and low risk groups, we employed the ssGSEA algorithm. Enrichment analysis was conducted to explore pathway differences between the high-risk and low-risk groups. Finally, the model genes were validated using an external validation set consisting of IPF cases, as well as single-cell data analysis.

resultsBased on machine learning algorithms, we constructed an ECM-related risk model. IPF patients in the high-risk group had a worse overall survival rate than those in the low-risk group. The model's AUC predictive values were 0.786, 0.767, and 0.768 for the 1-, 2-, and 3-year survival rates, respectively. The validation cohort validated these findings, demonstrating our model's effective prognostication. Chemokine-related pathways were enriched through enrichment analysis. Moreover, immune cell infiltration varied significantly between the two groups. Finally, the validation results indicate that the expression levels of all the model genes exhibited significant differential expression.

conclusionsBased on CST6, PPBP, CSPG4, SEMA3B, LAMB2, SERPINB4 and CTF1, our study developed and validated an ECM-related risk model that accurately predicts the outcome of IPF patients.