Fig. 1
Construction of the ICD risk model and its impact on LUAD patients survival. (A) NMF clustering heatmap divided the TCGA-LUAD cohort into clusters A and B. (B) Kaplan-Meier curves showed the survival outcomes of patients classified as A and B in TCGA-LUAD. (C) LASSO regression analysis was used to construct gene coefficients for the ICD risk score (IRS), where coefficients > 0 represent risk genes and coefficients < 0 represent protective genes. (D) The distribution of IRS across cluster A and cluster B. (E,F) Kaplan-Meier curves demonstrated the survival outcomes of HIRS and LIRS patients in the TCGA-LUAD. (E) Overall-survival (OS), (F) Progression-free survival (PFS). (G) Kaplan-Meier curves demonstrated the survival outcomes of HIRS and LIRS patients in the GEO-LUAD cohort (GSE10245, GSE14814, GSE41271, GSE42127, GSE68465, and GSE72094). (H) Forest plot showed that the IRS is an independent prognostic factor for lung adenocarcinoma patients across seven datasets, as determined by multivariate Cox regression analysis (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001).
