Influence of graft composition in patients with hematological malignancies undergoing ATG-based haploidentical stem...

To determine the influence of graft composition in haplo-HSCT, we summarized the long-term consequences of 251 consecutive transplantations from haploidentical donors. For donor-recipient HLA3/6-matched setting, 125 cases used G-CSF-mobilized BM and PBSCs mixtures, while 126 cases only used G-CSF-mobilized PBSCs in HLA4/6-matched transplantation. On the one hand, we wanted to explore the effect of harvests (CD34+ cells and TNCs dosages) on transplantation outcome in the context of haplo-HSCT no matter HLA4/6 or HLA3/6-matched setting. On the other hand, for patients using G-CSF-mobilized BM and PBSCs combination in HLA3/6-matched setting, we attempted to analyze whether TNCs or CD34+ cells from G-CSF-mobilized BM or G-CSF-mobilized PBSCs play the most paramount role on transplantation prognosis. Collectively, patients with hematologic malignancies receiving G-CSF-primed BM and PBSCs harvests had comparable consequences with patients only receiving G-CSF-mobilized PBSCs. Moreover, when divided all patients averagely according to the total amount of transfused nucleated cells, 3-year TRM of the intermediate group (13.06-18.05×10⁸/kg) was only 4.9%, which was remarkably reduced when compared to lower and higher groups with corresponding values 18.3%, 19.6% (P=0.026). The 3-year probabilities of OS and DFS of this intermediate group were 72.6% and 66.5%, which were slightly improved than the lower and higher groups. Most importantly, these data suggest that the transfused nucleated cells from G-CSF-primed BM above than 5.20×10⁸/kg could achieve remarkably lower TRM in haplo-HSCT receiving G-CSF-mobilized BM and PBSCs harvests. These encouraging results suggested that we could improve the efficacy of haplo-HSCT by adjusting the component and relative ratio of transfused graft cells. Nevertheless, the above findings should be confirmed in a randomized prospective comparative research with adequate follow-up.