Background: NK cells play a pivotal role in cancer immune vigilance. NK cell function is regulated by a balance between activating and inhibitory signals derived from various receptor-ligand interactions. Methods: 390 AML pts received 10/10 unrelated donor SCT. KIR genotype was performed by PCR with sequence specific primers. Hazard ratio (HR) for 2-year relapse was calculated using Fine-Gray regression and adjusted for disease risk index, remission status, pre-SCT MRD, conditioning regimen and presence of HLA-DP mismatch. KIR-ligand (K-L) match was defined as the presence a given KIR in the donor and the presence of its reported ligand in the patient (ex. 2DL1 and HLA-C2). KIRs that have no known ligands were not included in this analysis. The Table shows pt characteristics and K-L matches. Results: There was no correlation between the number of inhibitory or activating KIRs or KIR haplotype (A or B) and the probability of relapse after SCT. However, donor KIRs had a dramatic effect on relapse when they were considered together with the presence of the corresponding ligand in the pt. The 210 pts who had ≥3 inhibitory K-L matches had a significantly higher probability of relapse (HR=1.748, CI=1.147-2.667, p=0.009) than the remaining 180 pts. Similarly, the 96 pts who had at least one known activating K-L match had a lower probability of relapse (HR=0.581, CI=0.345-0.978, p=0.04). When we considered inhibitory and activating K-L matches together, we found that the 168 pts who had ≥3 inhibitory and no activating K-L matches had a significantly higher probability of relapse (HR 2.001, CI=1.376-2.908, p<0.001) than the 222 remaining pts. Conclusions: Donor-pt KIR-ligand matching should be taken into account when choosing unrelated donors for AML pts.