Optimal Timing of Allogeneic Stem Cell Transplantation (Allo-SCT) for Chronic Myeloid Leukemia (CML) Patients in the Tyrosine Kinase Inhibitor (TKI) Era

Introduction: TKIs have transformed the clinical history of chronic phase (CP) CML, largely replacing allo-SCT as the preferred first line treatment due to high efficacy and favorable safety profile. However some patients (pts) do not achieve an adequate response, develop resistance and/or transform to accelerated phase (AP) or blast phase (BP) CML. Response rates and duration are generally inferior beyond 2^nd TKI. Allo-SCT remains an important consideration for pts with TKI failure, but outcomes are worse in advance stage CML. Thus, timing of allo-SCT in the setting of suitable HLA-matched donor for CML pt with multiple failed TKIs is important to optimize outcome.

Methods: The aim of this study is to estimate residual life expectancy (RLE) for pts ages 18-65 with CP CML at initial diagnosis based on timing of allo-SCT in various settings: allo-SCT at CP1 CML, at CP2+ (after transformation to AP or BP), at AP, or at BP, or continuation of 2^nd or 3^rd TKI at CP1 CML, CP2+, AP or BP. The non-transplant cohort was extracted from MD Anderson (MDACC) Leukemia Department database, and the allo-SCT from Center for International Blood and Marrow Transplant Research (CIBMTR). MDACC cohort included patients initiating TKI between 1/1/2001-12/31/2013 and switched TKI due to failure as defined by European Leukemia Net (ELN) criteria; pts who underwent allo-SCT were excluded. CIBMTR cohort included pts diagnosed with CML between 1/1/2001-12/31/2013 and underwent HLA sibling matched (MRD) or unrelated donor (MUD) SCT. RLE was restricted to 10 yrs. Death due to any cause is considered an event.

Results: A total of 1361 pts were included for analysis: 138 in MDACC cohort and 1223 in CIBMTR cohort. In the MDACC cohort, 77 (56%) were male and median age was 46 years (yr) (range 18.5-64.4). In the CIBMTR cohort, 703 (57%) were male and median age was 39 yrs (range 18.2-64.8). In CP1 CML multivariate analysis (MVA) which included age (younger or older than 50 yrs) demonstrated an overall mortality hazard ratio (HR) for allo-SCT compared to no SCT of 2.4 [95% CI in () hereafter, (1.2-4.9; p=0.02)] with HR for MRD SCT of 1.9 (0.9-3.8, p=0.1) and HR for MUD SCT 3.3 (1.6-6.7, p=0.001), Table 1 and Figure 1. Mean RLE was 7.1 yrs (6.8-7.5) if patients underwent allo-SCT in CP1 CML vs. 8.5 yrs (7.6-9.5) if they did not, Table 2. Patients on their 1^st TKI had mean RLE of 7.0 yrs (6.6-7.5) if they underwent allo-SCT vs. 8.6 yrs (7.6-9.7) if they did not. HR for CP1 CML on their 1^st TKI was 2.6 (1.1-6.0; p=0.02) for allo-SCT vs. TKI continuation. HR was 2.3 (1.0-5.3, p=0.055) and 3.0 (1.3-6.9; p=0.01) for MRD and MUD SCT, respectively. Similarly, mean RLE was 6.8 yrs (5.6-8.3) for CP1 CML pts on 2^nd TKI if they underwent allo-SCT vs. 8.5 yrs (6.6-10) with TKI continuation. For CP1 CML pts on 2^nd TKI, HR for allo-SCT was 2.1 (0.5-9.2; p=0.3). For CP2+ CML pts, HR was 2.0 (0.8-4.9, p=0.13) for allo-SCT on MVA. HR was 1.8 (0.7-4.6, p=0.19) and 2.2 (0.9-5.4; p=0.10) for MRD and MUD SCT, respectively. Mean RLE was 4.7 (4.1-5.3) yrs for pts undergoing allo-SCT in CP2+ CML vs. 7.4 (5.6-9.7) yrs with TKI continuation. HR was 1.1 for allo-SCT at AP (0.6-2.1; p=0.8). Mean RLE was 5.3 (4.5-6.3) yrs for AP patients undergoing allo-SCT vs. 5.6 yrs (4.0-8.2) with TKI continuation. For those in BP, a trend towards higher survival with allo-SCT compared to no SCT was noted on MVA; HR 0.7 (0.5-1.1; p=0.099). HR=0.8 (0.5-1.3); p=0.33 for MRD SCT and HR=0.7 (0.4-1.0); p=0.08 for MUD SCT when compared to no SCT. Mean RLE was 2.6 (1.8-3.8) yrs in the pts who underwent allo-SCT in BP vs. 1.5 (0.9-2.8) yrs for those who continued their TKI.

Conclusion: Allo-SCT appeared to shorten survival for patients with CP1 CML but there was a trend towards benefit of SCT for those who progressed to BP CML. For CP2+ CML pts, allo-SCT trended towards worse survival. The role of allo-SCT in AP is less clear and should be determined on a case by case basis, dependent on mutation status, comorbidities, and the remaining TKI options available. However, given that this is not a clinical trial, it remains possible that patients referred for SCT may have had more aggressive disease so our results should be interpreted with caution due to the observational nature of the study and the small cohort of non-transplant pts. Considering the retrospective nature of this analysis, prospective randomized trials would be required to more formally define the optimal timing of SCT in different scenarios.

^contributed equally

View largeDownload slide

View largeDownload slide

Close modal