Comparable Long-Term Outcomes of Reduced-Intensity and Myeloablative Conditioning Allogeneic Hematopoietic Cell Transplantation By Minimal Residual Disease Kinetics during the Tyrosine Kinase Inhibitor-Based Chemotherapy Courses in Adults with Philadelphia-Chromosome Positive Acute Lymphoblastic Leukemia in First Remission

Background: The use of tyrosine kinase inhibitor (TKI)-based chemotherapy has demonstrated improved complete remission (CR) rates and increased applicability to allogeneic hematopoietic cell transplantation (HCT), thus allowing better survival in adults with Philadelphia chromosome-positive acute lymphoblastic leukemia (Ph-positive ALL). However, the sensitivity of Ph-positive ALL to reduced-intensity conditioning (RIC) versus myeloablative conditioning (MAC) by minimal residual disease (MRD) kinetics is not well established. Previously, we have confirmed that monitoring MRD kinetics is very important to predict long-term outcomes. Here, we examined a cohort of patients with Ph-positive ALL in CR1 and tried to compare the long-term outcomes of RIC-HCT vs MAC-HCT according to pre-HCT MRD kinetics.

Methods: During the period between 2000 and 2014, 173 adults (median age, 39 years [range, 16-65 years]) with Ph-positive ALL were included in this analysis. All patients received allogeneic HCT (68 RIC [fludarabine 150mg/m² + melphalan 140mg/m²] and 105 MAC [total body irradiation 13.2Gy + cyclophosphamide 120mg/kg]) in CR1 following two courses of first-line TKI (138 imatinib and 35 dasatinib)-based chemotherapy and had data on prospectively determined quantitative MRD kinetics. A total of 52 patients were excluded because of >CR1 pre-HCT status (n=26), transplants receiving umbilical cord blood grafts (n=14), and no TKI use before HCT (n=12). All but one RIC transplants received peripheral blood stem cells as a graft source (40 matched sibling donor, 11 matched unrelated donor, 17 mismatched unrelated donor), while MAC transplants used either bone marrow (n=73; 57 matched sibling donor, 8 matched unrelated donor, 6 mismatched unrelated donor) or peripheral blood stem cells (n=32; 2 matched sibling donor, 20 matched unrelated donor, 10 mismatched unrelated donor). The median time to transplant was 154 days (range, 119-291 days) in RIC transplants and 141 days (range, 112-280 days) in MAC transplants, respectively. Calcineurin inhibitors (cyclosporine for sibling donor transplants, tacrolimus for unrelated donor transplants) and methotrexate was used for graft-versus-host disease (GVHD) prophylaxis and antithymocyte globulin was administered to the patients who received mismatched unrelated donor grafts. If residual leukemia was detected in the absence of GVHD at 3 months after transplantation, calcineurin inhibitors were rapidly discontinued. MRD monitoring was centrally evaluated by real-time quantitative PCR (4.5 log sensitivity) through handling of bone marrow samples.

Results: After a median follow-up of 70.4 months (range, 16.0-176.8 mo), RIC regimen showed comparable 5-year cumulative incidence of relapse (CIR; 30.2% vs 27.9%, P=0.750), non-relapse mortality (NRM; 20.3% vs 15.5%, P=0.318), disease-free survival (DFS; 49.7% vs 56.6%, P=0.296), and overall survival (OS; 59.3% vs 62.1%, P=0.540) compared to MAC regimen. We further analyzed the impact of conditioning intensity on CIR and DFS according to MRD kinetics. Based on the MRD kinetics during the pre-HCT TKI-based chemotherapy courses, we classified patients into 3 subgroups: early-stable molecular responders (EMR, n=59), late molecular responders (LMR, n=57), and poor molecular responders (PMR, n=53). In all MRD-based subgroups of patients, no significant difference in CIR (EMR: 16.3% vs 6.2%, P=0.280; LMR: 10.5% vs 21.4%, P=0.334; PMR: 63.6% vs 59.4%, P=0.372) or DFS (EMR: 68.1% vs 78.1%, P=0.381; LMR: 49.6% vs 59.5%, P=0.369; PMR: 27.3% vs 34.2%, P=0.250) was observed between RIC and MAC. In multivariate analysis, LMR (HR, 2.36; 95% CI, 0.81-6.86; P=0.114) or PMR (HR, 9.05; 95% CI, 3.40-24.1; P<0.001) had higher relapse risk than EMR. Consequently, compared with EMR, LMR (HR, 2.02; 95% CI, 1.01-4.02; P=0.046) or PMR (HR, 3.79; 95% CI, 1.92-7.50; P<0.001) had poorer DFS. The presence of chronic GVHD, especially mild to moderate grade, was also independently associated with better DFS (HR, 0.28; 95% CI, 0.14-0.53; P<0.001). In addition, patients older than 40 years had higher risk of treatment failure in terms of NRM (HR, 4.25; 95% CI, 1.69-10.6; P=0.002).

Conclusion:RIC-HCT showed comparable long-term outcomes to MAC-HCT in all MRD-based subgroups of patients with Ph-positive ALL in CR1. Our data suggest that RIC-HCT is worthy of further investigation in prospective trials of adult Ph-positive ALL.