Fig. 1

Development of the lipid metabolism-based risk level approach for DLBCL patients. (A) Authentication of 523 lipid metabolism-related genes in three datasets (GSE181063, GSE10846, and NCICCR) using Venn diagrams. Changes in color denote differences in datasets. (B) LASSO coefficients of 16 obtained LMAGs over the 10-fold cross-validation approach. Vertical dotted lines denote the optimal values utilizing the minimum and 1-SE criteria. (C) Partial likelihood variance was uncovered using the LASSO regression model as well as the 10-fold cross-validation. Vertical dotted lines denote the optimal values utilizing the minimum and 1-SE criteria. (D) Forest plot of the linkages between the infiltrating levels of 16 prognostic molecules as well as the OS of the training cohort. The HR, 95% CI, and P-value were computed using univariate Cox regression analysis. (E) Coefficients for the 16 prognostic molecules within the Cox regression model. (F) The risk score distribution and survival levels of 16-gene signatures from the GSE181063 dataset. (G) Survival curves across the two risk groups from the GSE181063 dataset