Fig. 3

Association of L-scores with patient prognosis in lung adenocarcinoma data. A–D Kaplan-Meier curves of patients’ survival patterns in the TCGA LUAD [32] (A), the GSE8894 [39] (B), the GSE13213 [40] (C), and the GSE31210 [41] (D) dataset. Patients were stratified into two subgroups with high and low L-scores using the median as the cut-off values. E Forest plot showing the output from a multivariate Cox regression model for the dataset GSE31210 [41]. The covariates are binary values, including L-score, age, gender, and tumor stage. F, G Survival curves of the two subgroups stratified based on S-scores in the top 50% of patients with high L-scores (F) and the bottom 50% patients with low L-scores (G) in the dataset GSE31210 [41]. p-values were calculated by log-rank tests, and hazard ratios (HR) were determined by univariate Cox regression models