Allogeneic hematopoietic stem cell transplantation for the treatment of chronic myeloid leukemia in the era of tyrosine kinase inhibitors

The approval of imatinib in 2001 has changed the landscape of CML management.1–3 With its excellent efficacy in the International Randomized Study of Interferon and STI571 (IRIS), imatinib has become the standard of care for newly diagnosed patients. This change challenges the previous role of allogeneic hematopoietic stem cell transplantation (allo-HSCT) in CML treatment. However, allo-HSCT still plays an important role in the treatment of CML patients because of the following reasons: (1) up to one-third of patients in the first chronic phase (CP1) of CML (CML-CP1) are resistant to or intolerant of imatinib therapy;4,5 (2) patients with the T315I mutation are highly resistant to all tyrosine kinase inhibitors (TKIs); (3) pediatric and young patients with an identical sibling donor have a strong willingness to be cured; (4) patients with advanced disease have poor response to TKIs; and (5) the outcomes of HSCT have improved in recent years. So in this era, the indication was more complex and stratified in more detail, based on weighing the disease related risk and SCT related risk. HSCT FOR PATIENTS WITH CML IN CP1 In the patient group with optimal responses to TKIs, overall survival (OS) was projected to be close to 100% after six to seven years in CP1 patients;4,5 hence, newly diagnosed CML-CP1 patients were recommended imatinib as first-line therapy instead of HSCT. Consequently, the annual number of allo-HSCTs in CML-CP1 has fallen markedly in the last decade. An European Group for Blood and Marrow Transplantation (EBMT) study showed that the HSCTs yearly for CML is 1396 in 1999, 802 in 2004, and 434 in 2007.2,3 In China, proportion of allo-HSCT for CML patients decreased, with a rate of 29.6% in 2007 to 23.8% in 2008 to 18.2% in 2009.6 However, the subgroup of patients in CML-CP1 are considered to be benefitted from HSCT. PATIENTS WITH FAILURE OF TKIS TREATMENT By directly targeting Bcr-Abl kinase, imatinib actually leads to long-lasting cytogenetic remissions, which in turn, improves survival, but many patients develop resistance, or are intolerant of imatinib because of side effects. In the IRIS trials, 15%-25% of patients had failed imatinib mesylate (IM) therapy because of drug intolerance or resistance, which occurred at a rate of 4% per year.4,5 CML patients in CP who failed to respond to imatinib eventually develop poor outcomes. Marin et al7 reported that the outcome in the CML-CP patients with imatinib failure at +3 months, +6 months, +12 months, and +18 months were worse than those without imatinib failure at the same time point. The corresponding OS rates for with vs. without imatinib failure were 60.2% vs. 93.2%, 81.8% vs. 95.5%, 87.1% vs. 95.1%, and 87.8% vs. 98.5%, whereas the progression-free survival (PFS) rates were 56.2% vs. 84.6%, 73.4% vs. 87.1%, 76.0% vs. 90.0%, and 76.4% vs. 97.1% respectively. Therefore, patients who satisfy the criteria for IM failure should be switched to HSCT or second-generation TKIs.8,9 Some of the patients with IM failure get more benefits from early HSCT than the second-generation TKIs. There are various methods by which these subgroup of patients could be identified. Breccia et al10 report the response of patients who previously experienced IM failure to second-generation TKIs can be predicted accurately using the Hammersmith score. The Hammersmith score includes three factors: the cytogenetic response (CyR) to IM, the Sokal score, and recurrent neutropenia on IM. Based on the presence of these three factors, patients were categorized into good risk (n=24), intermediate risk (n=27), or poor risk (n=29). The respective 2.5-year cumulative incidence of achieving complete cytogenetic response (CCyR) after the initiation of second-generation TKIs were 100%, 52.2%, and 13.8% respectively, for these three risk groups. Patients with high Hammersmith scores may be candidates for early allo-HSCT, especially if they have a low EBMT score. Jabbour et al11 found that the response to second-generation TKIs among patients in the CP1 stage of IM failure can accurately be predicted using mutational scoring based on their in vitro median inhibitory concentrations (IC50). It can also be predicted through mutation analysis. Mutations of the BCR-ABL1 kinase domain (KD) were described in 42%-90% of the patients with TKI resistance.12 Philadelphia-positive (Ph+) patients who harbor imatinib-resistant Bcr-Abl kinase domain mutations have higher likelihood of developing additional mutations associated with resistance to second- or third-line TKIs; thus, these patients may benefit from early HSCT.13 Patients with mutations were more likely to develop advanced disease states, with worse outcomes after HSCT in the case of those who progress to the advanced stage. Jabbour et al14 reviewed the outcomes of imatinib- resistant CML patients: CP (n=34), accelerated phase (AP, n=9), and blast phase (BP, n=4), who underwent HSCT and BCR-ABL1 sequencing. Mutations occurred in 40% of the patients (n=19); 78.9% of the patients (15 out of 19) were in the advanced stage. At the time of imatinib failure, 69% of the mutant patients and 35% of the non-mutant patients progress into the AP or BP (P=0.03). Up to 42 patients (89%) responded to HSCT, whereas 32 (68%) had at least a major molecular response (MMR). The 2-year event-free survival (EFS) rates were 36% and 58% (P=0.05) for the mutant and non-mutant groups respectively. The 2-year OS was 44% and 76% (P=0.02) respectively. Therefore, patients in the CP1 stage with mutations should undergo early HSCT, just in case imatinib resistance occurs. For patients switched to second-generation TKIs after failure to IM, secondary failure might occur. Hence, the response to next generation-TKI should be monitored closely and evaluated early. Allo-HSCT remains the only therapeutic option for patients with second-generation TKI failure as well as with the first generation TKI. Markiewicz et al15 reported that 48 CML patients, who failed previous treatment with TKI (imatinib, 37 patients; imatinib with dasatinib, five patients; imatinib, dasatinib, and nilotinib, three patients; dasatinib, two patients; and imatinib and nilotinib, one patient), received allo-HSCT from HLA-matched siblings (n=15) or from 10/10 (n=21) and 9/10 (n=12) match unrelated donors (MUD). The 5-year estimated OS was 79%. The data implies that myeloablative allo-HSCT is a feasible and effective curative therapy for TKI-resistant CML. Additionally, patients with suboptimal response to TKIs have worse outcome than the non-suboptimal group. Marin et al7 reported that the outcome in the suboptimal group after +6 months, +12 months, and +18 months were worse than those of the non-suboptimal groups. Comparing the suboptimal vs. the non-suboptimal groups, the +12 month OS rate were 85.4% vs. 98.4%, whereas the PFS rates after +6 months and +12 months were 61.5% vs. 91.4% and 73.4% vs. 96.1% respectively. But whether early HSCT is better than TKIs for patients in the CP stage with suboptimal response is yet to be established. HSCT FOR PATIENTS WITH THE MUTATION (T315I) Mutations were found in 25% of CP patients (12/47), 33% of AP (5/15), and 71% of BC patients (5/7). The TKI-resistant T315I mutation was identified at high mutation frequencies among TKI-exposed cases, rather than among newly diagnosed patients or among hydroxyurea-exposed patients.16,17 The BCR-ABL/T315I mutation is highly resistant to all available TKIs. Thus, HSCT is the first line and the only salvage therapy if patients harbor the BCR-ABL/T315I mutation. Velev et al18 reported on HSCT for eight patients with CML who were resistant to TKIs and had the T315I mutation. At the time of HSCT, two patients were in CP1, three patients in AP, and three patients in the second chronic phase (CP2). The best response after HSCT were complete molecular response (CMR) in three patients, CCyR in four patients, and complete hematological response (CHR) in one patient, whereas no response was observed in one patient. The best outcome was for patients who underwent HSCT in CP1; both remained alive and in CMR after +14 months and +42 months. After a median of 13 months from HSCT, five patients were alive, three patients were in CMR, one patient was in CCyR, and one patient was in CHR. Therefore, HSCT is an effective strategy for patients with CML who have the T315I mutation, particularly in the early stages. PEDIATRIC/ADOLESCENTS YOUNG ADULT PATIENTS WITH MSD CML, with an annual incidence of approximately one per million children and adolescents younger than 20 years, is even rarer in children younger than age 4.19,20 The superiority of imatinib over all-HSCT is not well established in pediatric patients. Only a few studies have addressed the outcome of bone marrow transplantation (BMT) in children with Ph+ CML. In a registry report from the EBMT, the OS after three years for children transplanted in CP1 from sibling donors or from unrelated donors was 75% and 65% respectively,21–24 which is not inferior to imatinib. In a report25 the outcomes of 31 children (≤ 18 years) were diagnosed with CP-CML from 2002 to 2008 were retrospectively analyzed. All received imatinib at 260 300 mg/m2 per day. The 2-year major cytogenetic response (MCR) rate was 82% and complete cytogenetic response equivalence (CCRe) was 70%. The 5-year PFS and OS rates were 68% and 75.8% respectively. Limited data show that the outcomes of HSCT and imatinib are equally good on their 5-year follow-up, and both are better tolerated by children and adolescents.21–25 Thus, the decision to treat with either TKI therapy or transplantation is controversial.21–25 However, if the follow-up times are extended, HSCT should have better outcomes than TKIs. Additionally, children and adolescents are treated longer with TKIs. Therefore, possible toxicities may make long-term TKI therapy less attractive. Allo-HSCT as a potential cure is more attractive, with HSCT techniques improving over time. A large, well-designed study involving pediatric patients in CP1 is needed to evaluate the role of all-HSCT in the TKI era. For young adult patients with CML, IM is the standard therapy according to the European LeukemiaNet (ELN) and National Comprehensive Cancer Network (NCCN) guidelines,26,27 but some data show that the outcomes of HSCT in certain patients is equally good. Excellent outcomes post sibling identical donor (SID) HSCT for CML patients with CP1 were reported in some single arm studies.28–30 Radich et al30 used an intravenous BuCy preparative regimen with busulfan targeted to achieve a steady-state plasma concentration of ≥900 ng/ml. This approach resulted in an estimated 3-year OS and disease-free survival (DFS) rate of 86% and 78% respectively. Huang et al29 reported that for 89% of 3-year OS and LFS in SID HSCT for CML patients, the 3-year OS and DFS reached 91% when patients with low EBMT scores. For patients in CP1 who are unexposed to imatinib with an identical sibling donor, a regimen of allo-HSCT has been evaluated in a study that compared allo-HSCT and IM. Jiang et al31 reported a prospective cohort study comparing the outcomes of allo-HSCT from an HLA-matched sibling donor for CML in CP1 with imatinib. Two hundred and nine patients were assigned to the allo-HSCT group (median age: 38 years) and 254 patients were assigned to the imatinib group (median age: 37 years). For the patients in the early CP (ECP, a CML duration <12 months), allo-HSCT was inferior to imatinib with a 6-year EFS of 74.3% vs. 90.3% (P=0.001) and 6-year OS rates of 80.2% vs. 99.4% (P <0.001); allo-HSCT was not inferior to imatinib, with 6-year PFS rates of 95.9% vs. 97.3% (P=0.303). For those in the LCP, both treatments resulted in similar EFS and OS rates, with a probability of more than 90% for the 6-year EFS rate, and a 6-year OS rate of approximately 80%, HSCT is superior to imatinib with a 6-year PFS rate of 100% vs. 86.0% (P=0.035). The report shows that allo-HSCT confers significant PFS advantages and is not inferior to imatinib as the first-line therapy for patients with imatinib-naïve CML in the LCP. Although allo-HSCT in carefully selected, younger patients may have a similar OS to that of IM-treated patients,31 the age of patients at diagnosis, the availability of well-matched sibling donors, and the variability of outcomes in the centers should be considered. HSCT FOR PATIENTS WITH CML IN ACCELERATED PHASE Allo-HSCT cases with advanced CML were not reduced in the era of TKI, because more HSCT shifted to the advanced stage. Imatinib could create a window of treatment through downstage of the patients. A large study by the Center for International Blood and Marrow Transplant Research (CIBMTR) that included 2436 patients documented an increase in HSCTs for second CP (CP2) or AP from 24% to 41% among all transplants for CML1 and from 32% to 53% of the annual transplantations in the DRST study.32 For patients presenting in AP at diagnosis, imatinib is the standard care prescribed by the NCCN guidelines. In a single arm study, for imatinib-treated patients in AP (n=75), the cumulative hematological response (HR) rate was 93.3%, a complete HR (CHR) rate of 85.3%, and a return of CP2 rate of 8% after a median follow-up of 23 months. The estimated 4-year PFS and OS rates were 48.2% and 52.2%, respectively.33 Secondary failure and mutation is the problem. Patients who can benefit from early HSCT may be defined by predictive factor pre-TKI treatment or response during TKI treatment. Kantarjian et al34 reported that pretreatment anemia and a lack of CyR after 3 months on imatinib therapy as negative predictors of survival in AP-CML. In that study, the estimated 4-year survival rates were 88% for low-risk patients (i.e., those with no negative factors present) and 60% for other patients.34 Early HSCT should be considered for patients who are expected to have a poor response to TKIs. Our data shows that high and intermediate-risk AP-CML patients get more benefits from allo-HSCT than from imatinib. Jiang et al35 compared the outcomes of imatinib (n=87) versus allo-HSCT (n=45) for AP-CML in a cohort study. A multivariate analysis of the total population revealed that a CML duration ≥12 months, hemoglobin <100 g/L, and peripheral blood blasts ≥5% were independent adverse prognostic factors for both OS and PFS. It categorized patients into low-risk (no factor), intermediate-risk (any one factor), and high-risk (at least two factors) groups. The low-risk patients have similar survival rates, the EFS and OS of intermediate-risk patients are not significantly improved, but allo-HSCT is considerably superior to imatinib among high-risk patients. Therefore, allo-HSCT confers significant survival advantages for high- and intermediate-risk patients with AP-CML compared with imatinib treatment. However, in low-risk patients, the outcomes of allo-HSCT and imatinib treatment are equally good. Therefore, both are acceptable treatment options. Transplantation for those classified as low-risk AP may be delayed as a salvage option at the time of entering the high- and intermediate-risk groups, or at the time of being classified as low-risk when there is limited imatinib. HSCT can be undertaken for patients who have alternative donors other than matched unrelated donor. In 2008, a research reported by Huang et al36 showed HLA-mismatched or haploidentical HSCT without in vitro TCD for CML improved the outcomes of patients in AP and BC. The probability of a 4-year OS was 73.3% in the AP group. Haploidentical donors HSCT for advanced CML patients are acceptable in experienced centers. Should HSCT be undertaken for patients with CML-AP who are not eligible for myeloablative preparative regimens due to old age? This should be considered individually. A study from EBMT used a variety of RIC regimens for 186 patients receiving RIC, with 2-year OS in AP of 24%.37 Many new regimens were studied recently and the Flu/BU/ATG-based RIC regimens appeared to have lower TRM.37–39 Although allo-HSCT could be the preferred as a second-line option among intermediate- or high-risk patients AP patients, whether it was superior to the second-generation TKIs for all the AP-CML patients after imatinib failure remains to be determined. HSCT FOR PATIENTS WITH CML IN BLASTIC CRISIS PHASE The prognosis of patients with CML in BC remains dismal even with the availability of the BCR-ABL TKI (imatinib).33,40 Jiang et al33 reported 49 cases of CML patients in the BP treated with 400 mg or 600 mg of imatinib once daily. The cumulative HR rate was 63.3%, the rate of returning to chronic phase (RCP) was 18.4% in a median of 4.5 months. The cumulative major cytogenetic response (MCyR) and CCyR rates were both 12.2%. For patient HR, the estimated 1/2-year PFS and OS rates are 32.8%/15.8% and 46.0%/21.0% respectively. It only offers short-term benefits in most cases. In recent years, a vast majority of CML blastic crisis (BC) cases arise in patients already on imatinib-based treatment; in TRIS, resistance to IM develop in 90% of BP within 4 years.4,5 Second-generation TKIs also give a poor response. Saussele et al41 reported 24 cases of imatinib failure after loss of CP (AP, n=10; BC, n=14). The 3-year OS after the start of second-generation TKI treatment is 19% for resistant patients with advanced disease. For patients presenting BP at diagnosis, imatinib followed by allo-HSCT is the standard care under the NCCN guidelines. For patients progressing to AP or BP after imatinib failure, second-generation followed by allo-HSCT is the standard care. Allo-HSCT seems to be a viable option for BC-CML patients who have attained remission. Comparative analysis of CML patients with BC showed that the efficacy of transplantation is significantly better than that of TKIs. In a comparative cohort trial performed by Huang et al (data unpublished), the 3-year OS was 7% in the TKI group, and 45% in HSCT group (P <0.05). For patients with BC CML, myeloablative allo-HSCT combined with imatinib in CP2 can achieve better outcomes after HSCT. Luo et al42 reported myeloablative allo-HSCT combined with imatinib for advanced CML in 15 patients with AP (n=6) or BC (n=9). Up to 11/15 (73.3%) achieved CHR with pre-transplant imatinib, 6 (40%) achieved CyR. The early TRM was only 6.7%. The 5-year estimated rates of relapse, TRM and OS were 21.0%±10.8%, 13.7%±10.8%, and 66.0%±12.4%, respectively; up to of the 15 patients with With BC or BC CML, myeloablative allo-HSCT similar outcomes after HSCT. et retrospectively reviewed all patients who received an for CML in CP2 at from to after had or BC. The 3-year OS was for and for BC The disease donor stem cell or cytogenetic response to HSCT or complete versus not were found in the progress has been in SCT haploidentical HSCT for patients with hematological for haploidentical HSCT have been reported in a few centers. et treated patients with allo-HSCT, who achieved CP2 through imatinib after the of BC. patients received HSCT from haploidentical After a median follow-up of 24 months six of patients and are in CCyR, one of relapse, the of versus disease Patients transplanted with bone marrow from haploidentical donors had high A research reported by Huang et al36 showed that HLA-mismatched or haploidentical HSCT without in vitro TCD for CML improves the outcomes of patients in the advanced stage. The probability of 4-year OS was 61.5% in the BC group. TRM is not with a TRM at and TRM at For patients with an unrelated donor is available in limited time. Haploidentical donors are better alternative because all patients can a haploidentical donor without and for donor are alternative is It has many potential advantages over HSCT, of availability and the of to the donors, a lower risk of and chronic versus disease as well as and A recent analysis patients with CML who received the 2-year OS rates were and 32% among AP, and BP patients A study on is still It has the of limited cell number and of was recently for adult patients and no data is available on CML patients. For patients and those with who be candidates for myeloablative HSCT, RIC regimens HSCT to et reported that CML patients with advanced disease were were treated with RIC After a median follow-up of years, the 5-year OS and PFS were 33% and respectively. After years, and years, the were and respectively. However, in a study from EBMT of 186 patients that received a variety of RIC the 2-year OS in BC was The poor outcome after RIC SCT in BC that the may not be strong and additional is For patients with BC, IM is used as a to HSCT, and HSCT should be undertaken as as possible in cases when CP2 HSCT has less in CML treatment than during the pre-TKI era. However, it is still a important alternative therapy for many CML patients. It could become the preferred second-line option after the failure of the first-line TKI It is an alternative therapy for patients presenting in the AP or BP at diagnosis, or pediatric or young patients with a matched sibling donor. It is first-line treatment for patients with the T315I mutation, for all patients with failed second- generation TKI for patients progressing to AP or BP after imatinib should to develop new that to the and the of cure for patients with CML.