Transplants of unrelated cord blood or sibling allogeneic peripheral blood stem cells/bone marrow in adolescent and young adults with chronic myeloid leukemia: comparable outcomes but better chronic GVHD-free and relapse-free survival among survivors with cord blood

Adolescent and young adult (AYA) patients with hematological malignancy aged 15 to 39 years are recognized as a separate entity, and the efficacy and safety of unrelated cord blood transplantation (CBT) for chronic myeloid leukemia (CML) in AYA patients has not been reported. From March 2002 to June 2015, total of 106 CML patients received allogeneic hematopoietic cell transplantation (allo-HCT) in our center. Included in the present study were CML patients aged 15 to 39 years who received unrelated CBT or sibling allo-HCT, and 74 consecutive AYA patients with CML enrolled in this analysis. The day-100 cumulative incidences of grade 2–4 aGVHD and grade 3–4 aGVHD were similar following CBT and sibling-PBSCT/BMT. The cumulative incidences of cGVHD and extensive cGVHD were 21.7% and 5.3% in the CBT cohort, which were significantly lower than those in the sibling-PBSCT/BMT cohort (58.0% and 45.5%), respectively (p = 0.046, 0.008). There was no significant difference between the two cohorts in terms of transplant-related mortality (TRM), relapse, and long-term survival (overall survival and leukemia-free survival). The 5-year probability of GVHD-free/relapse-free survival (GRFS) was 47.9% and 33.4% in the CBT and the sibling-PBSCT/BMT cohorts, respectively (p = 0.632); among patients who survived more than 100 days after transplantation (n = 61), the 5-year probability of chronic GVHD-free, relapse-free survival (CRFS) was 66.2% in the CBT cohort, which was significantly higher than that in the sibling-PBSCT/BMT cohort (37.4%) (p = 0.037). Our study suggests that for AYA patients with CML, transplantation using unrelated CB offers comparable outcomes to sibling -PBSCT/BMT, including similar aGVHD, TRM, relapse, and long-term survival; however, from the perspective of quality of life, unrelated CBT have a lower incidence of cGVHD and a higher CRFS among survivors.


INTRODUCTION
Tyrosine kinase inhibitors (TKIs) have remarkably improved the clinical outcomes of patients with chronic myeloid leukemia (CML), but little attention has been paid to the adolescent and young adult (AYA) group. AYA patients with hematological malignancy aged 15 to 39 years are recognized as a separate entity, with unique features of their medical and psychosocial needs which require age-appropriate treatment and care. Two retrospective clinical studies [1,2] indicated that AYA patients with CML treated with TKIs had significantly lower complete cytogenetic and molecular response, and inferior event-free survival as compared to older patients.

Clinical Research Paper
Although the number of allogeneic hematopoietic cell transplantation (allo-HCT) in CML has been dramatically decreased nowadays, it remains the only curative option for patients with intolerance or resistance to TKIs, or with advanced stages (accelerated phase or blast crisis, AP or BC). HLA-identical allo-HCT from a sibling donor is considered as the standard transplant model for CML, but at least two thirds of patients eligible for transplantation can not find such a donor. The development of cord blood (CB) registries and the increasing number of CB units collected allow more frequently using CB as an alternative graft source; the advantages are the immediate availability of donor cells, absence of donor risk, a lower risk of transmitting infections, and decreased graft-versus-host disease (GVHD) with preserved graft-versus-leukemia effects. Several previous reports [3][4][5][6][7] have demonstrated that unrelated cord blood transplantation (CBT) can be regarded as a reasonable option for CML patients requiring allo-HCT but lacking a suitable sibling donor. However, the efficacy and safety of unrelated CBT for CML in AYA patients was not reported. In this study, we retrospectively analyzed the outcomes of CML in AYA patients receiving unrelated CBT compared with those patients receiving sibling allo-HCT, with an emphasis on transplant-related complications and long-term survival, in order to explore possible survival advantages of CBT.

DISCUSSION
Several retrospective clinical studies [3][4][5][6][7] had investigated the role of unrelated CBT for the treatment of CML, and these results indicated that unrelated CBT could be used as a reasonable alternative for CML patients who needed transplantation but lacked a suitable donor. Until now, very few clinical studies have focused on the efficacy and safety of allo-HCT for CML in AYA patients. A recent CIBMTR cohort analysis [8] evaluated the outcomes of myeloablative HCT in children and    young adults with CML-CP, and indicated that HLAmatched sibling donor and source of BM led to the best outcomes compared to unrelated donor and PBSC; however, this study did not cover the graft source of cord blood, and the impact of CBT on the outcomes of children and young adults CML was still a unsettled issue. In the current study, we firstly demonstrated that, for AYA patients with CML, unrelated CBT had similar incidence of aGVHD, TRM and relapse, and similar long-term survival (OS, LFS) compared to sibling-PBSCT/BMT. Moreover, it is worth mentioning that unrelated CBT had a lower incidence of cGVHD and a higher GRFS rate among patients who survived for more than 100 days after transplantation. Our data also indicated that unrelated CBT was associated with delayed neutrophil and platelet recovery compared to sibling-PBSCT/BMT. This might be due to the insufficient number of TNC and CD34+ cells in the CB graft. Nevertheless, we did not observe more bacterial or fungal infections in the CBT cohort owing to delayed neutrophil engraftment. Several strategies have been designed to accelerate neutrophil recovery, such as transplant of double CB units, injection of CB into BM, co-infusion of mesenchymal stem cells, or cytokine-mediated ex vivo expansion, and achieved an improvement in neutrophil or platelet engraftment. We observed that transplantation with CB was associated with a very lower incidence of cGVHD compared to that of sibling donor (21.7% vs 58.0%, p = 0.046; extensive cGVHD: 5.3% vs 45.5%, p = 0.008). Numerous recent studies demonstrated that transplantation of CB had lower incidence and severity of cGVHD than that of related or unrelated donors. Gutman et al [9] demonstrated a significantly lower incidence of moderate to severe cGVHD following double CBT (8%) as compared with PB transplant from matched unrelated donor (44%) in hematological malignances (p = 0.0006). Our previous study of advanced CML (AP/ BC) [4] indicated that patients receiving CBT had slightly lower incidence of cGVHD as compared to patients receiving allo-PBSCT/BMT (19.5% vs 39.6%, p = 0.09). Furthermore, we also found that, for AML patients [10], CBT had a significantly reduced rate of cGVHD (13.7% vs 28.3%; p = 0.047) or extensive cGVHD (9.9% vs 24.1%; p = 0.039) compared with that of MSD. This phenomenon may be associated with 10-fold fewer T cells existed in CB, and these T cells are mostly with a naive phenotype characterized by atypical functional properties and little baseline cytotoxicity [11]. On the other hand, we speculated that more patients in the sibling cohort receiving PBSC (or PBSC plus BM) (n = 42, 89.4%) as a graft source may contribute to the high incidence of cGVHD, since T-cell amounts in the PBSC are higher than those in the BM or CB. Adding ATG in the conditioning to deplete T cells in vivo may be one of the approaches to decrease the morbidity and mortality of cGVHD following PBSCT, and a recent report from Europe [12] showed that inclusion of ATG in the conditioning resulted in a significantly lower rate of cGVHD and a higher rate of GRFS after myeloablative HLA-identical sibling PBSCT. However, other investigators found that additional ATG use might increase the disease relapse and transplantrelated infections [13][14][15][16].
Among long-term relapse-free survivors after transplant, health related quality of life (HRQoL) of post-transplantation is a great concern for AYA CML patients. Most reported studies have shown that HRQoL correlates with the incidence and the severity of cGVHD, and the extensive cGVHD would have a profound negative impact on HRQoL [17][18][19]. GRFS is now a new composite endpoint of transplantation in current clinical trials [20], which focused on severe aGVHD, cGVHD requiring systemic treatment (extensive cGVHD), TRM, or relapse; therefore, GRFS represents a comprehensive measure of HRQoL after allo-HCT. Our previous data illustrated that [10], for AML patients, transplantation with CB had similar rates of TRM and severe aGVHD but less cGVHD and a lower risk of relapse, which translated into better GRFS as compared with sibling donor. In the present study, we found that, among patients who survived more than 100 days after transplantation, the 5-year probability of CRFS in the CBT cohort was significantly higher than that in the sibling-PBSCT/BMT cohort (66.2% vs 37.4%) (p = 0.037), and this indicated that AYA survivors of CML who received unrelated CBT had a better HRQoL without ongoing morbidity and experienced real recovery after transplantation.
Although German CML Study IV indicated that CML-CP patients receiving allo-HSCT had the similar survival compared with that of matched patients receiving TKIs [21], we do not recommend allo-HSCT as firstline therapy in CML-CP nowadays. Transplant should be reserved for those who are resistant or intolerant to at least one second generation TKI, or those who are in AP or BC with suitable donors. Our comparison suggests that for AYA patients with CML, transplantation using unrelated CB offers comparable outcomes to sibling -PBSCT/BMT, including similar aGVHD, TRM, relapse, and long-term survival; in addition, from the perspective of HRQoL, unrelated CBT has a lower incidence of cGVHD and a higher CRFS among survivors. However, some limitations were obvious in this study. First, this was only a retrospective study, and multivariate analyses did not performed in this study due to the small number of patients in each group. Second, the disease entity was heterogeneous, such as the fact that CML-CP and CML-AP or BC patients were mixed in this study, and more patients in the CBT cohort were associated with advanced stages (AP or BC) at first diagnosis. Third, we could not estimate the impact of post-HCT TKI use on the outcomes of transplantation, because TKI intervention was not planned for those patients before transplantation. www.impactjournals.com/oncotarget

Patient eligibility
From March 2002 to June 2015, total of 106 CML patients received allo-HCT [40 of unrelated CBT, 65 of sibling allo-HCT, and 1 of unrelated PBSCT] at Anhui Provincial Hospital (32 were previously reported [4]). Included in the present study were CML patients aged 15 to 39 years who received unrelated CBT or sibling allo-HCT. Seventy-four consecutive AYA patients with CML enrolled in this analysis, which included 27 patients receiving unrelated CBT and 47 patients receiving sibling allogeneic PBSCT or bone marrow transplantation (sibling-PBSCT/ BMT) (27 received PBSC plus BM, 15 received PBSC, and 5 received BM). The baseline patient related characteristics were showed in Table 1. There were no significant differences between the CBT and the sibling-PBSCT/ BMT cohorts in terms of patient age, sex, pre-transplant treatment, ECOG performance status, cytomegalovirus (CMV) serology, and EBMT score before transplant. There were more patients with advanced stages (AP or BC) at first diagnosis in the CBT cohort (n = 17, 63.0%) than in the sibling-PBSCT/BMT cohort (n = 11, 23.4%) (p = 0.003); however, at the time of transplantation, number of patients with advanced stages were similar between two cohorts [8 (29.6%) in the CBT cohort, and 5 (10.6%) in the sibling-PBSCT/BMT cohort] (p = 0.19).

Transplant characteristics
HLA-identical sibling allo-HCT was the first selection. However, if the patient had no suitable sibling donor (HLA-identical or 1-antigen-mismatched), or there was not sufficient time to wait an unrelated donor, unrelated CBT would be performed. CB units that were serologically matched for at least 4 of 6 HLA loci and which contained a minimum count of 3 × 10 7 ⁄ kg total nucleated cells (TNC) and/ or 1.2 × 10 5 ⁄ kg CD34+ cells of the recipient weight before freezing. Patients without enough single cord blood unit were considered double unrelated CBT. Nineteen out of 27 patients (70.4%) received single-unit CBT, and the other 8 patients (29.6%) received double-unit CBT.
GVHD prophylaxis was composed of cyclosporine (CSA) and mycophenolate mofetil (MMF) in all transplantations as previously described [4,22,23], with the exception of 4 patients in the sibling-PBSCT/BMT cohort also receiving additional short-term methotrexate (MTX).

Definitions and statistical analyses
The definitions of neutrophil and platelet engraftment, primary graft failure, acute GVHD and chronic GVHD, transplant-related mortality (TRM), relapse, overall survival (OS), and leukemia-free survival (LFS) were previously published [22][23][24][25][26]. Patient-, disease-, and transplant-related variables, such as categorical variables were measured using χ 2 test, and continuous variables were measured using Mann-Whitney U test between the CBT and sibling-PBSCT/BMT cohorts. The variables selected for analysis were age, gender, disease stage at diagnosis or in transplant (chronic phase or advanced stages), pre-transplant treatment, reasons for transplant, ECOG performance status, recipient CMV serology, EBMT score, donor to recipient gender, HLA match, ABO compatibility, conditioning regimens (myeloablative conditioning or reduced intensity conditioning), GVHD prophylaxis, TNC dose, and CD34+ cell dose, and post-transplantation TKIs use. The probabilities of engraftment, GVHD, TRM, and relapse were estimated by the cumulative-incidence function method with considering competing risks. The end point of OS was death of any cause, and the end point of LFS was relapse or death; the end points of GVHD-free/ relapse-free survival (GRFS) were severe aGVHD (grade 3-4 aGVHD), extensive cGVHD or chronic GVHD requiring systemic treatment, relapse, or death. The probabilities of OS, LFS, and GRFS were generated by the Kaplan-Meier method. Statistical analyses were conducted using R statistical software (R Foundation for Statistical Computing, Vienna, Austria). Differences with р< 0.05 were considered significant.