Fig. 4
From: Impacts of heavy smoking on non-coding RNA expression for patients with esophageal carcinoma
Construction of a smoking-related prognostic model in patients with esophageal carcinoma. A The LASSO Cox regression analysis was constructed from 19 candidates derived from univariate Cox regression, and the tuning parameter (λ) was calculated based on the partial likelihood deviance with tenfold cross-validation. Then, an optimal log λ value is indicated by the dotted line in the plot. B The ROC curve for validation of the efficiency of the smoking-related risk score. C Kaplan‒Meier analysis with log-rank test for survival analysis in smoking-related subgroups. The median was defined as the cutoff value. D The risk plot of survival time and risk score. E The proportion of survival status in smoking-related subgroups