Figure 3.

Impact of donor KIR2DS1 and KIR3DL1 combinations on relapse incidence, overall survival, and nonrelapse mortality. The graphs show the outcome of patients who were grouped according to their donors’ KIR2DS1 status (activating vs nonactivating) and KIR3DL1/HLA-B subtype combinations (strong inhibiting vs noninhibiting/weak inhibitory) in terms of relapse incidence, overall survival, and nonrelapse mortality. Patients were classified according to the algorithm published by Boudreau et al.¹⁹ The P values represent the score tests of univariate Cox regression models.