Research Paper Volume 13, Issue 19 pp 22947—22962

A novel prognostic signature based on immune-related genes of diffuse large B-cell lymphoma

The process of variable selection. (A) Forest map of 26 candidate immune-related genes selected by univariate Cox regression analysis associated with DLBCL overall survival in the discovery cohort. (B, C) The performance of least absolute shrinkage and selection operator (LASSO) analysis. (D) LASSO and random forest analysis intersecting genes were selected. (E) Forest map of multivariate Cox regression analysis to establish a prognostic signature.

Figure 1. The process of variable selection. (A) Forest map of 26 candidate immune-related genes selected by univariate Cox regression analysis associated with DLBCL overall survival in the discovery cohort. (B, C) The performance of least absolute shrinkage and selection operator (LASSO) analysis. (D) LASSO and random forest analysis intersecting genes were selected. (E) Forest map of multivariate Cox regression analysis to establish a prognostic signature.