The number of NK cells in the stem cell product is protective against disease relapse. Relationship between the number of 4 different cell subsets administered within the stem cell infusion and clinical outcome. The median number of CD34⁺ cells, lymphocytes, T cells, and NK cells was defined, and transplants were divided into those where the infusion contained a cell dose either above (blue) or below this median (red) value. Clinical outcome measures were RI (A), aGVHD (B), NRM (C), and OS (D). This demonstrates that NK cell dose is the only parameter that influences disease relapse (P < .0001). Cumulative incidence analysis was tested using the Gray method. Log-rank tests were used to analyze the Kaplan-Meier survival curves.