Significance of CD34

Reduced toxicities with reduced-intensity and non-myeloablative conditioning (collectively, non-MAC) regimens have expanded access of allogeneic hematopoietic cell transplantation (HCT) to older and less fit adults with acute myeloid leukemia (AML). Relative to myeloablative conditioning (MAC), non-MAC regimens rely more heavily on immunological effects to prevent post-HCT relapse from AML. Whether there is an optimal CD34+ cell dose in this setting remains controversial, and how measurable residual disease (MRD) might impact the interplay between CD34+ cell dose and outcome is unknown. To evaluate the association between infused CD34+ cell doses and post-HCT outcomes, including relapse, overall survival, and graft-versus-host disease (GVHD), in adults with AML undergoing non-MAC allografting, using pre-HCT MRD status as a modulating factor. In this retrospective single-center study, we identified 455 adults with AML or MDS/AML in first or second morphologic remission who underwent non-MAC and received mobilized peripheral blood stem cell allografts from 10/10-HLA matched (n = 352), HLA mismatched (MMUD; n = 80), or HLA-haploidentical (n = 23) donors and had available flow cytometric data on infused CD34+ cell doses. Pre-HCT MRD testing by 10-color MFC was available for all patients. Probabilities of relapse, overall survival (OS), and GVHD were analyzed. In multivariable analysis of the 10/10-HLA matched cohort, the lowest CD34+ dose tertile (T1) was associated with a significantly higher relapse (HR = 2.01 [1.29 to 3.14], P = .002), but decreased moderate-to-severe chronic GVHD (HR = .55 [.34 to .89], P = .014) compared to the intermediate dose (T2). In subgroup analyses, the 3-year relapse incidence was significantly higher in low (T1) versus high (T2-3) CD34+ cell doses in the MRDneg cohort (42% versus 26%, P = .018), and was numerically higher, but not statistically different in the MRDpos cohort (60% versus 48%, P = .60). However, there was not a significant interaction between MRD status and CD34+ cell dose group (P = .51). There was no significant impact of low (T1) versus high (T2-3) CD34+ cell dose on 3-year overall survival (49% versus 58%, P = .12), 100-day noon-relapse mortality (3% versus 4%, P = .50), 28-day neutrophil (97% versus 98%, P = .11) and platelet (92% versus 95%, P = .20) engraftment, or 100-day grade 3 to 4 acute GVHD (8% versus 8%, P = .98). In exploratory analysis of the MMUD cohort, although the lowest CD34+ tertile was not associated with relapse (HR = 1.12 [.35 to 3.53], P = .85), there was a nonsignificant trend toward increased relapse risks with higher (T3) versus intermediate (T2) CD34+ doses (HR = 2.50 [.88 to 7.15], P = .087). There was no association between CD34+ cell doses and relapse in the HLA haploidentical cohort. In patients with AML undergoing 10/10-HLA matched non-MAC allografting, lower CD34+ cell doses were associated with an increased relapse risk-particularly in the MRDneg cohort-and lower GVHD risk, with no impact on OS. These findings suggest that selection of "optimal" CD34+ cell dose targets may need to be tailored to individual patient's risk of disease relapse versus chronic GVHD. Future studies are needed to better evaluate these findings across alternative HCT platforms (including MMUD and haploidentical cohorts) and to further define the interaction between graft cell composition, pre-HCT disease burden, and conditioning intensity.