Early Tm/Term ratio as a prognosticator of OS. (A) UMAP from Abbas et al,⁴ colored by CD8⁺ T-cell subsets, and split by group (HD, NonRes, and Res). (B) Point-range plot from Abbas et al,⁴ showing the pairwise (Res vs NonRes) proportional difference (Monte Carlo/permutation test) for each CD8⁺ T-cell subset. The colors indicate the statistical significance (red, FDR < 0.05; blue, FDR ≥ 0.05); the vertical dashed lines mark the absolute value of log2 fold difference (FD) cutoff for significance. (C) Kaplan-Meier estimates of OS in patients (BeatAML2 cohort, n = 444 patients) with a Early Tm/Term ratio score above (n = 222) and below (n = 222) the median value. Survival curves were compared using a log-rank test. (D-E) Box plots showing the Early Tm/Term ratio (y-axis) in relation to ELN 2017 risk score (D) and the NPM1, RUNX1, ASXL1, and TP53 mutations (x-axis) (E) computed from patients in the BeatAML cohort. (F) Forest plot of diagnosis features (patient age and ELN 2017) and Early Tm/Term score associated with survival in multivariate Cox proportional hazard analyses. ADV, adverse; INT, intermediate; FAV, favorable; HR, hazard ratio for death; WT, wild type.