Fig. 7

Predictive modeling and survival analysis selecting prognostic markers. (A) Lasso coefficient profiles of prognostic markers across a range of lambda values in CLL. (B) Partial likelihood deviance plot for model tuning across different λ values. (C) Risk model constructed based on markers of Lasso model. (D) Receiver operating characteristic curves evaluating the model’s discriminative ability. AUC, area under the curve. (E) Kaplan-Meier survival plots stratified by low and high-risk groups calculated from the risk model. (F) A prognostic nomogram integrating the expression of selected markers to predict the probability of survival at 12, 36, and 60 months. (G) Calibration plots demonstrating the accuracy of the nomogram’s predictions compared to actual observed outcomes.