Prognostic Factors for Outcomes of Allogeneic HSCT for Children and Adolescents/Young Adults With CML in the TKI Era

The breakthrough effects of tyrosine kinase inhibitors (TKIs) have lessened indications for allogeneic hematopoietic stem cell transplantation (HSCT) in chronic myeloid leukemia (CML). However, HSCT is still attractive for children and adolescents/young adults (AYAs) requiring lifelong TKI therapy. Nevertheless, little has been reported on the outcomes of large clinical studies of HSCT targeting these age groups. This study aimed to identify prognostic factors for the outcomes of HSCT, including reduced-intensity conditioning (RIC)-HSCT, for children and AYAs with CML in the TKI era. We performed a registry analysis for 200 patients with CML aged <30 years who underwent pretransplant TKI therapy from the observational nationwide database established by the Japanese Society for Transplantation and Cellular Therapy. The patients received bone marrow (BM), peripheral blood (PB), or cord blood (CB) from either related or unrelated donors. The indication for HSCT for individual patients was determined by the institution according to European LeukemiaNet recommendations and other guidelines. The 5-year overall survival (OS) rates for patients with chronic phase (CP) (n = 124), accelerated phase (AP) (n = 23), and blastic phase (BP) (n = 53) at diagnosis were 82.8%, 71.1%, and 73.3%, respectively, with no significant difference (P =.3293). The strongest predictor of engraftment was transplant source, with CB (hazard ratio [HR], 0.33) and PB (HR, 2.00) (compared with BM) being independent unfavorable and favorable predictors, respectively. Transplant source was also an independent predictor of chronic GVHD, with PB (HR, 1.81) and CB (HR, 0.39) (compared with BM) being unfavorable and favorable predictors, respectively. The strongest predictor of OS rate for patients with CP at diagnosis was disease phase at HSCT, with second or greater CP, AP, or BP (HR, 2.81) (compared with first CP [CP1]) being an unfavorable predictor. In addition, patients with CP at diagnosis who had major and complete molecular responses at HSCT had excellent outcomes, with 5-year OS rates of 100% and 94.4%, respectively. The 5-year OS rate was compared between RIC (n = 31) and myeloablative conditioning (MAC) (n = 58) in patients with CP1, both of which were 89.3%, with no significant difference (P = .9440). On univariate analysis for the RIC cohort with CP at diagnosis, the age at HSCT (HR, 1.27) (increase per year) and the time from diagnosis to HSCT (HR, 1.83) (increase per year) were significant predictors for OS. Our study demonstrates that RIC may be an appropriate alternative to MAC for children and AYAs with CP1. As for the transplant source, we recommend first selecting BM because of a higher engraftment rate compared to CB and a lower incidence of chronic GVHD compared to PB. Although HSCT in the status of a major molecular response is desirable, it is not advisable to continue TKI pointlessly long because age at HSCT and timing of HSCT are prognostic factors that determine survival. The decision to perform RIC-HSCT instead of continuing TKI should be carefully made, considering the possibility of transplant-related complications.