Figure 3.
Defining LIRS. (A) Workflow of machine learning pipeline identifying the most prognostic inflammatory proteins. (B) Kaplan-Meier survival curves of patients with high, medium, and low LIRS risk separated by score tertile in the training cohort. (C) Cubic spline showing the log HR across the continuous LIRS score in the training cohort. (D) Kaplan-Meier survival curves of patients in the validation cohort separated by LIRS high-, medium-, and low-risk groups by score tertile. (E) Forest plot of the multivariable Cox model in validation cohort incorporating significant prognostic factors for OS in AML with LIRS. For categorical variables ELN and LIRS, ELN favorable and low LIRS group were used as reference level, respectively.