- Email: [email protected]
Myeloablative versus Reduced-Intensity Conditioning in Patients with Myeloid Malignancies: A Propensity Score-Matched Analysis
Accepted Manuscript Title: Myeloablative Versus Reduced-Intensity Conditioning in Patients with Myeloid Malignancies: a Propensity Score Matched Analysis. Author: Hassan Sibai, Umberto Falcone, Uday Deotare, Fotios V. Michelis, Jieun Uhm, Vikas Gupta, John Kuruvilla, Jeffrey H. Lipton, Matthew D. Seftel, Hans A. Messner, Dennis (Dong Hwan) Kim PII: DOI: Reference:
S1083-8791(16)30327-5 http://dx.doi.org/doi: 10.1016/j.bbmt.2016.08.030 YBBMT 54385
To appear in:
Biology of Blood and Marrow Transplantation
Received date: Accepted date:
7-7-2016 31-8-2016
Please cite this article as: Hassan Sibai, Umberto Falcone, Uday Deotare, Fotios V. Michelis, Jieun Uhm, Vikas Gupta, John Kuruvilla, Jeffrey H. Lipton, Matthew D. Seftel, Hans A. Messner, Dennis (Dong Hwan) Kim, Myeloablative Versus Reduced-Intensity Conditioning in Patients with Myeloid Malignancies: a Propensity Score Matched Analysis., Biology of Blood and Marrow Transplantation (2016), http://dx.doi.org/doi: 10.1016/j.bbmt.2016.08.030. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Myeloablative versus reduced-intensity conditioning in patients with myeloid malignancies: a propensity score matched analysis. Hassan Sibai*, MD, Umberto Falcone*, MD, PhD, Uday Deotare, MD, Fotios V. Michelis, MD, PhD, Jieun Uhm, MD, PhD, Vikas Gupta, MD, John Kuruvilla, MD, Jeffrey H. Lipton, MD, Matthew D. Seftel, MD, Hans A. Messner, MD, and Dennis (Dong Hwan) Kim, MD, PhD. Allogeneic Blood and Marrow Transplant Program, Princess Margaret Cancer Centre, University Health Network, University of Toronto, Toronto, Canada. *Both these authors contributed equally to this work.
Corresponding author: Dennis (Dong Hwan) Kim, MD, PhD Rm 5-126, Department of Medical Oncology and Hematology Princess Margaret Cancer Center, 610, University Ave., Toronto, Canada M5G 2M9 Tel: 1-416-946-4501 x 2464 Fax: 1-416-946-4563 Email: [email protected] Presented in part at the 56th American Society of Hematology Annual Meeting in San Francisco, California, December 6-9, 2014.
Short title: Propensity score analysis of allo-HSCT conditioning regimens.
Financial Disclosure Statement: The authors state that they have no conflict of interests
1 Page 1 of 22
HIGHLIGHTS RIC is comparable to MAC for AML/MDS PSM analysis balances pretransplant patient characteristics Addition of low-dose TBI to allo-SCT conditioning could reduce the risk of relapse
Abstract Reduced-intensity conditioning (RIC) has been shown to have similar overall survival (OS) but higher relapse rates when compared to myeloablative (MAC) regimens in patients with myeloid malignancies. Using propensity score matching (PSM) analysis, well-balanced pairs of different variables can be compared effectively. We retrospectively compared allogeneic hematopoietic stem cell transplantation patients with acute myeloid leukemia and myelodysplasia receiving RIC regimen (FBT 200; fludarabine 30mg/m2/day for 4 days, busulfan 3.2 mg/kg/day for 2 days, and total body irradiation (TBI) 200 cGy) or MAC regimen (FBT 400; fludarabine 50 mg/m2/day for 4 days, busulfan 3.2 mg/kg/day for 4 days and TBI 400 cGy). Two hundred and forty-eight patients (127 MAC, 121 RIC) were included in the analysis. No statistical difference was observed in 2-year OS (RIC 45.2±5.0%, MAC 51.7±5.2%; p=0.541), non-relapse mortality (NRM, RIC 28.7±2.8% MAC 34.7±4.6% p=0.368), and acute (p=0.171) or chronic (p=0.605) GVHD at one year between FBT200 and FBT400. Cumulative incidence of relapse at 2 years (CIR) was statistically different (RIC 26.1±2.6%, MAC 14.2±3.5%; p=0.033). When PSM was applied to the study population, 42 case-control pairs were evenly matched. PSM analysis confirmed no statistical difference in 2-year OS (RIC 49.0±9.1%, MAC 54.9±7.7%; p=0.718), NRM (RIC 22.2 ± 2.3%, MAC 33.3±2.8%;
2 Page 2 of 22
p=0.238). CIR did not differ statistically between FBT200 and FBT400 (RIC 25.7±2.6%, MAC 9.5±1.1%; p=0.315) in the PSM pairs. Our study shows that after applying PSM, FBT 200 RIC conditioning is able to give OS, NRM, and CIR rates comparable to FBT 400 MAC conditioning.
Keywords: Propensity score matching; allogeneic stem cell transplant; conditioning regimen; acute myeloid leukemia; myelodysplasia.
3 Page 3 of 22
Introduction Allogeneic hematopoietic stem cell transplantation (allo-HSCT) is a curative treatment
modality for patients with acute myeloid
leukemia (AML)
and
myelodysplastic syndrome (MDS) as well as other hematological disorders. 1-4 Myeloablative conditioning (MAC) regimens, such as high doses of total body irradiation (TBI) plus chemotherapy2 or combination chemotherapy regimens including busulfan and cyclophosphamide1 have been widely used while the use of reduced intensity conditioning (RIC) has been increasing over the past 2 decades.7-9 RIC includes reduced dose of chemotherapy and/or low dose TBI. Adding low dose TBI to RIC regimens has been shown to be a well-tolerated option with anti-leukemic activity comparable to MAC regimens.17 However it is still in debate whether RIC results in similar post-transplant outcomes compared to MAC. While suspected to increase the risk of relapse due to lower intensity of chemotherapy, RIC is associated with less conditioning-related toxicities and non-relapse mortality (NRM) when compared to MAC.5-7, 10-11 In addition to recipient age, other factors including disease status at transplant, donor type, cytogenetics, hematopoietic cell transplantation-comorbidity index (HCT-CI)
12
,
contribute to the outcome variables in both MAC and RIC transplants. In most transplant studies, these pre-transplant factors are not entirely matched between comparable groups. The propensity score is the probability of treatment assignment conditional on observed baseline characteristics. Propensity score matching (PSM) analysis is used to balance the variables affecting the choice of a treatment among different treatment options.13, 14 We conducted a retrospective study including patients undergoing allo-HST for AML or MDS, with the aim to compare MAC and RIC conditioning regimens, using PSM
4 Page 4 of 22
analysis to obtain well matched pairs of patients. Our hypothesis is that by this mean, OS and relapse rate obtained with RIC are comparable to those obtained with MAC.
Patients and Methods Study design and eligibility criteria We retrospectively compared MAC and RIC conditioning regimens in patients who underwent allo-HST for AML or MDS at Princess Margaret Cancer Centre (PMCC) between January 2009 and December 2013. This study was approved by the Research Ethics Board of the University Health Network/Princess Margaret Cancer Centre. Only AML patients in first (CR1) or second complete remission (CR2) and MDS patients with <10% bone marrow blasts at the time of the transplant were included in the study. For AML, cytogenetic risk at diagnosis was characterized as Favorable/intermediate versus Adverse, as described by SWOG/ECOG.15 In patients with good cytogenetic risk allo-HST was offered for AML refractory to first-line treatment, or at relapse after achieving CR2. For MDS cytogenetics, 2008 WHO MDS cytogenetic minimal criteria were used.16 Inclusion was restricted to patients receiving Fludarabine/Busulfan (FB) plus TBI with either RIC (Fludarabine 30mg/m2/day for 4 days, Busulfan 3.2mg/kg/day for 2 days and TBI 200 cGy) or MAC conditioning (Fludarabine 50mg/m2/day for 4 days, Busulfan 3.2mg/kg/day for 4 days and TBI 400 cGy). The decision to offer RIC was primarily based on patient’s age (≥60 years) and/or the presence of comorbidities.5, 20-22
Stem cell source consisted of matched-related (MRD), matched-unrelated
(MUD), or mismatched unrelated (MMUD) donors. Donor peripheral blood stem cells (PBSC) were mobilized with G-CSF. Haploidentical or cord blood allograft recipients were excluded from this study. Graft-versus-host disease (GVHD) prophylaxis in 5 Page 5 of 22
MRD allo-HSCT consisted of cyclosporine A (CSA) combined with either mycophenolate mofetil (MMF, 15 mg/kg by mouth or IV twice daily (dose was rounded to the nearest multiple of 250) from day 0 for 30 days and then stopped without taper) or methotrexate (15 mg/m2 on HSCT day +1 and 10 mg/kg on HSCT days +3, +6, and +11). In patients receiving MUD allo-HSCT, T-cell depletion was performed using low-dose alemtuzumab or rabbit antithymocyte globulin (ATG) in combination with CSA. Pre-transplant variables included in the PSM were: age, HCT-CI, remission status (CR1 vs CR2, and MDS with <10 % blasts) at the time of transplant, diagnosis (AML vs MDS), cytogenetic risk group (adverse-risk vs favorable (relapsed) and intermediate risk), donor type (MRD vs MUD and MMUD), and period effect (year of transplant). We adopted PSM analysis to adjust the risk factors affecting the choice of conditioning regimens by doing well balanced pairs of RIC and MIC patients.
Definitions of statistical endpoints Overall survival (OS) duration was measured from the date of allo-HSCT until death from any causes. Alive patients were censored on the date of their last follow-up. Cumulative incidence of relapse (CIR) was calculated from the date of allo-HSCT until relapse. Relapse was defined as ≥5% blasts in a bone marrow aspirate or peripheral blood, or the development of extra-medullary leukemia following alloHSCT. Non-relapse mortality (NRM) was calculated as death without evidence of disease relapse. Acute and chronic GVHD were diagnosed and graded using established criteria. 18-19
Statistical analysis
6 Page 6 of 22
Patient and disease characteristics were reported using descriptive statistics. Data were updated as of December 2013.The main outcome variables of interest included: OS, cumulative incidence of relapse (CIR) and cumulative incidence of NRM. CIR and NRM are competing risks setting. For relapse incidence, death from non-relapse cause was accounted as a competing event, while relapse was taken account for competing event for NRM. Cumulative incidence of grade II-IV acute GVHD and chronic GVHD at one year were estimated taking into account death as a competing event. In order to adjust for any potential biases derived from imbalanced pretranplant factors between FBT200 vs FBT400 groups, we have adopted propensity score matching analysis. For an initial step, propensity score was calculated initially using binary logistic regression model. The following 7 independent pre-transplant factors were included into the binary logistic regression model for calculation of propensity score: age, remission status, diagnosis, cytogenetic risk group (adverserisk vs favorable/intermediate), donor type, HCT-CI and period effect. Then each patient from FBT200 group was matched and paired to another patient from FBT400 treatment group with the lowest differences of propensity score using in-house script. For risk factor analysis, we have included the 7 covariates previously described with conditioning regimen (i.e. MAC vs RIC) in the analysis for OS, CIR and NRM. These analyses were performed repetitively in the whole cohort of patients and in the PSM groups, separately. After the univariable analysis, multivariable analysis adopted Cox’s hazard proportional regression model for OS and Fine-Grey method for CIR and NRM. Stepwise selection procedure was applied using the criteria for variable selection, p=0.05 for variable entry and p=0.1 for variable removal. In addition to all the variables remained in the multivariable model, the variable of interest which is conditioning regimen is included in the final model as shown in Table 3 and 4. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated for the significant risk factors. Paired analysis was adopted throughout the PSM analysis for survival. For MVA in PSM groups, only the variables statistically significant (plus 7 Page 7 of 22
conditioning regimen) at MVA for the whole cohort were considered. Statistical analyses were performed using EZR on R Commander Version 1.11.24
Results Patients, diseases and treatment characteristics Patient characteristics are summarized in Table 1. A total of 248 patients (male: 134; 54%) were transplanted using PBSC as a graft source. Median age at transplant for the entire cohort was 54 years (range: 18-71). FBT400 was administered to 127 patients (51.2%) and FBT200 to 121 (48.8%). Remission status was CR1 in 132 patients (53%) while 56 patients (23%) underwent transplant in CR2. Adverse risk cytogenetics was present in 54 patients (21.7%) while patients with good and intermediate risk cytogenetics were 194 (78.3%). Source of stem cells was MRD in 103 cases (41.5%), while MUD and MMUD were 115 (46.4%) and 30 (12.1%), respectively. Demographics and pre-transplant characteristics were imbalanced between FBT200 and FBT400 groups. Older age (P<0.001), higher HCT-CI score (p=0.001) and more MRD (p=0.061) were observed in the RIC group. However, no differences were observed between FBT200 and FBT400 groups in CR status at HSCT (p=0.316), subtype of diagnosis (AML vs MDS, p=0.206), or cytogenetic risk group (p=0.261). Using PSM analysis to overcome baseline imbalances, 42 case-control pairs (84 patients) were selected (Table 1). Pre-transplant variables became well balanced between the two groups (FBT400 and FBT200) after applying PSM. Median age at transplant remained different between the 2 PSM groups (55 years in FBT400, 58 years in FBT200; p=0.009).
8 Page 8 of 22
Overall outcomes With a median follow-up of 18 months among survivors in the overall population (n=248), the 2-year OS, NRM and relapse incidence rate was 48.0±3.6%, 34.6±3.6% and 24.8±3.5%, respectively. There was no difference in OS (45.2±5.0% in RIC vs 51.7±5.2% in MAC patients at 2 years; p=0.541) or NRM (28.7±2.8% in RIC vs 34.7±4.6% in MAC patients at 2 years; p=0.368) between FBT200 and FBT400 groups. The CIR at 2 years was higher in the RIC group (26.1±2.6% in RIC vs 14.1±1.4% in MAC; p=0.033, Figure 1A). In the PSM subgroup of patients, no statistical difference was noted in 2-year OS (RIC 49.0±9.1%, MAC 54.9±7.7%; p=0.718), NRM (RIC 22.2±2.3%, MAC 33.3±2.8%; p=0.238), or CIR (RIC 25.7±2.6%, MAC 9.5±1.1%; p=0.315, Figure 1B) between FBT200 and FBT400 groups.
Acute and chronic GVHD according to conditioning for overall and PSM pairs The overall incidence of grade 3-4 acute GVHD was 19.4±1.9%. When comparing between the two conditioning regimens, FBT400 and FBT200, there was no difference in rates of acute and chronic GVHD of any grade (Table 2). PSM pairs showed no difference for acute (any grade; FBT400 71.4% vs FBT200 52.4%, p=0.143) or chronic (any grade; FBT400 39.3% vs FBT200 64.8%, p=0.138) GVHD rate at one year (Table 2).
Univariate and Multivariate analysis of prognostic factor for transplant outcomes Overall Survival
9 Page 9 of 22
In the whole population, age ≤60 years (p=0.014), CR1 (p= 0.007), diagnosis of AML (p= 0.015), related donor (p=0.002) and favorable/intermediate risk cytogenetics (p= 0.016) were all predictors of better OS at UVA. Only age (p=0.01), disease state at HSCT (p=0.018) and cytogenetic risk group (0.027) remained significant at multivariate analysis (MVA). Conditioning regimen was statistically not significant at UVA (p=0.541). In PSM pairs, UVA for the same variables (Table 3) showed no differences in OS for the 2 conditioning regimens (all patients: 48.0±3.6 months, FBT400: 54.7±7.7 months, FBT200 49.5±9.1 months; p=0.718). Statistically significant variables were donor type (p=0.002), diagnosis (p=0.006) and cytogenetic risk group (p<0.001). At MVA, only cytogenetic risk group remained statistically significant, with adverse risk group
showing
a
worse
OS
(15.6±9.1
months
vs
66.4±6.4
months
in
favorable/intermediate risk; p=0.002).
Non-relapse mortality In the overall population of patients, disease status (p=0.004), donor type (p=0.001), and diagnosis (p=0.001) were significant at UVA. No impact of conditioning regimen was observed on NRM even when PSM was applied. At MVA only donor type (p=0.002) and diagnosis (p=0.004) remained significant (Table 4). In PSM pairs, CR2 disease state at transplant (p=0.008), MMUD type (p=0.003), AML diagnosis (p=0.005), and adverse cytogenetic risk group (p=0.024) were related to higher NRM at 2 years.
Transplant from MRD (vs MUD/MMUD) resulted in
improved NRM at MVA {HR=1.031; 95% CI [1.085-4.236]; p= 0.028}.
10 Page 10 of 22
Cumulative incidence of relapse UVA for the whole patient cohort showed that CIR at 2 years was adversely affected by age >60 years (p=0.021) and RIC regimen (vs MAC; HR=1.932; 95% CI [1.0873.433]; P=0.033). At MVA, CIR was higher in patients aged >60 years (p=0.009), not in CR1 (p=0.020), with MDS diagnosis (p=0.006), and adverse risk cytogenetics (p=0.027). Conditioning regimen did not remain statistically significant at MVA {FBT200: HR=1.341, 95% CI [0.662-2.714]; p=0.401} (Table 5). For CIR in PSM pairs, only adverse cytogenetic risk group remained significant at both UVA and MVA (Table 5).
Discussion A previous study at a single institution comparing a MAC regimen including IV busulfan plus cyclophosphamide (ivBUCY) with two RIC regimens of fludarabine plus IV busulfan 6.4 mg/kg (FB2) or IV busulfan 12.8 mg/kg (FB4) in patients with AML and MDS at various stages of disease, showed no difference in 2-year OS (ivBUCY 50%, FB2 47%, FB4 49%; p=NS).5 Kharfan-Dabaja et al. reviewed FB2 and FB4 regimens in patients with AML in CR1 undergoing allo-HSCT at various participating centers from European Bone Marrow Transplant Group (EBMT) showing that FB2 can be safely used in these patients when aged ≥ 50 years while it does not affect OS when used in patients <50 years.6
11 Page 11 of 22
Recently, randomized BMT CTN 0901 study was stopped prematurely because RIC resulted in higher relapse rates and lower transplant related mortality (TRM) compared to MAC, with a statistically significant advantage in RFS for patients receiving MAC; however the right choice of conditioning intensity remains challenging with a necessity to balance the risk of relapse with TRM.25 In our study of 248 patients with AML or MDS, we found that with a median follow-up of 18 months there was a higher incidence of relapse in the RIC group. However, since there was a fundamental difference of age criteria between the 2 groups, with MAC used for patients ≤60 years and RIC for pts >60 years, this analysis was further subjected to PSM. After balancing the variables, we found 42 case/control pairs between the 2 groups. We were able to conclude that there were no statistical differences in OS, NRM, or relapse rate. An interesting finding was that chronic GVHD rates at 1 year were 64.8% vs 39.3% in RIC and MAC arms, respectively. Although statistical significance was not reached, a trend towards higher rates of chronic GVHD in RIC patients was observed. Since the incidence of chronic GVHD is associated with reduced relapse rates, we postulated that unevenness in the initial findings was effectively highlighted in the PSM analysis. PSM analysis has not been frequently used as a statistical method for comparison of groups in allo-HSCT studies. PSM is able to match pairs in an efficient manner across groups. While previously PSM has been used in this setting for patients with acquired aplastic anemia undergoing allo-HSCT 23, to the best of our knowledge, this study is the first to use PSM to compare conditioning regimens (RIC vs MAC) for alloHCT in hematological malignancies, in a sizable number of patients. It is realistically impossible to design a prospective clinical trial randomizing MRD vs alternative donor for allo-HSCT in aplastic anemia. However, by adopting PSM, they
12 Page 12 of 22
have used the score by comparing 25 patients in each group of MRD, MUD, and MMUD transplants from a total of 225 acquired aplastic anemia patients, thus balancing imbalanced variables using PSM.23 In our study, negation of significant variables was found so that PSM is needed as an equalizer in multiple variables in allo-HSCT studies and hence deserves due recognition in the current world. In our study we were able to group-match age, disease risk status, donor type, diagnosis, conditioning regimen, cytogenetic risk group, comorbidity score (HCT-CI), and period of transplant. When we compared our results of FB plus low dose TBI to previous experiences with FB without radiation 6, we found that the addition of low-dose TBI was able to reduce the relapse rate at 2 years: 9.5% our study (PSM), 23% Kharfan-Dabaja et al) although it is indirect comparison. This was seen with both MAC and RIC regimens, although the relapse rates were higher in RIC as compared to MAC. Previously, Russell et al.17 showed that the addition of low dose TBI (400cGy) to the FB regimen was able to reduce the relapse risk (HR= 0.29) as well as the remaining parameters, including OS and DFS. Hence our data suggest that the addition of low-dose TBI to FB combination could reduce the risk of relapse. In addition to the inherent limitations of the retrospective nature of the study, the patient comparable group was small. By applying PSM the overall sample size came down from 248 to 84. The pre-transplant mutational status (i.e. nucleophosmin-1 (NPM1) mutation, FMS-related tyrosine kinase-3 (FLT-3) internal tandem duplication mutation), chemotherapy treatment for the myeloid malignancy, GVHD prophylaxis regimen, and CMV status at the time of transplant were not taken into consideration at the time of analysis. In conclusion, our study shows that RIC conditioning using combination of FB+TBI (200cGy) is able to give similar survival rates and CIR rates compared to FBT400
13 Page 13 of 22
MAC conditioning. Outside of the propensity model, the CIR at 2 years was higher in the RIC group. This conditioning regimen will require further studies and evaluation in the years to come.
Acknowledgements: Previously presented in part at the 56th American Society of Hematology Annual Meeting in San Francisco, California, December 6-9, 2014. Conflict of Interest: The authors declare no conflict of interest. Authorship Contributions: HS and UF contributed equally to this work. HS, UF, DK, UD analyzed and interpreted data and drafted the paper. HS, JU, and FM collected the data. HS, DK, UF, UD: Concept and designed the study. HS, DK performed the statistical analysis. HS, DK, UF, UD, FM, JU, VK, JK, JL, MS, and HM critically revised the manuscript.
References [1] Santos GW, Tutschka PJ, Brookmeyer R, Saral R, Beschorner WE, Bias WB et al. Marrow transplantation for acute nonlymphocytic leukemia after treatment with busulfan and cyclophosphamide. N Engl J Med. 1983;309(22):1347-1353. [2] Blume KG, Beutler E, Bross KJ, Chillar RK, Ellington OB, Fahey JL et al. Bonemarrow ablation and allogeneic marrow transplantation in acute leukemia. N Engl J Med. 1980;302(19):1041-1046. [3] Deeg HJ. Hematopoietic cell transplantation for myelodysplastic syndrome. Am Soc Clin Oncol Educ Book. 2015:e375-80. [4] Brown JR, Kim HT, Armand P, Cutler C, Fisher DC, Ho V et al. Long-term followup of reduced-intensity allogeneic stem cell transplantation for chronic lymphocytic leukemia: prognostic model to predict outcome. Leukemia. 2013;27(2):362-9. [5] Shimoni A, Hardan I, Shem-Tov N, Yeshurun M, Yerushalmi R, Avigdor A et al. Allogeneic hematopoietic stem-cell transplantation in AML and MDS using myeloablative versus reduced-intensity conditioning: the role of dose intensity. Leukemia. 2006;20(2):322-328. 14 Page 14 of 22
[6] Kharfan-Dabaja MA, Labopin M, Bazarbachi A, et al. Comparing i.v. BU dose intensity between two regimens (FB2 vs FB4) for allogeneic HCT for AML in CR1: a report from the Acute Leukemia Working Party of EBMT. Bone Marrow Transplant. 2014;49(9):1170-1175. [7] Slavin S, Nagler A, Naparstek E, Kapelushnik Y, Aker M, Cividalli G et al. Nonmyeloablative stem cell transplantation and cell therapy as an alternative to conventional bone marrow transplantation with lethal cytoreduction for the treatment of malignant and nonmalignant hematologic diseases. Blood. 1998;91(3):756-63. [8] Socie G, Clift RA, Blaise D, Devergie A, Ringden O, Martin PJ et al. Busulfan plus cyclophosphamide compared with total-body irradiation plus cyclophosphamide before marrow transplantation for myeloid leukemia: long-term follow-up of 4 randomized studies. Blood. 2001;98(13):3569-3574. [9] Reshef R, Porter DL. Reduced-intensity conditioned allogeneic SCT in adults with AML. Bone Marrow Transplant. 2015;50(6):759-769. [10] Alyea EP, Kim HT, Ho V, Cutler C, Gribben J, DeAngelo DJ et al. Comparative outcome of nonmyeloablative and myeloablative allogeneic hematopoietic cell transplantation for patients older than 50 years of age. Blood. 2005;105(4):18101814. [11] Gooley TA, Chien JW, Pergam SA, Hingorani S, Sorror ML, Boeckh M et al. Reduced mortality after allogeneic hematopoietic-cell transplantation. N Engl J Med. 2010;363(22):2091-2101. [12] Sorror ML, Maris MB, Storb R, Baron F, Sandmaier BM, Maloney DG et al. Hematopoietic cell transplantation (HCT)-specific comorbidity index: a new tool for risk assessment before allogeneic HCT. Blood. 2005;106(8):2912-2919. [13] Austin PC. An Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Studies. Multivariate Behav Res. 2011;46(3):399424. [14] Rosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects. Biometrika. 1983;70(1):41-55. [15] Slovak ML, Kopecky KJ, Cassileth PA, et al. Karyotypic analysis predicts outcome of preremission and postremission therapy in adult acute myeloid leukemia: a Southwest Oncology Group/Eastern Cooperative Oncology Group study. Blood. 2000;96(13):4075-4083. [16] Vardiman JW, Thiele J, Arber DA, et al. The 2008 revision of the World Health Organization (WHO) classification of myeloid neoplasms and acute leukemia: rationale and important changes. Blood. 2009;114(5):937-951. [17] Russell JA, Irish W, Balogh A, et al. The addition of 400 cGY total body irradiation to a regimen incorporating once-daily intravenous busulfan, fludarabine, and
15 Page 15 of 22
antithymocyte globulin reduces relapse without affecting nonrelapse mortality in acute myelogenous leukemia. Biol Blood Marrow Transplant. 2010;16(4):509-514. [18] Przepiorka D, Weisdorf D, Martin P, et al. 1994 Consensus Conference on Acute GVHD Grading. Bone Marrow Transplant. 1995;15(6):825-828. [19] Pavletic SZ, Lee SJ, Socie G, Vogelsang G. Chronic graft-versus-host disease: implications of the National Institutes of Health consensus development project on criteria for clinical trials. Bone Marrow Transplant. 2006;38(10):645-651. [20] McClune BL, Weisdorf DJ, Pedersen TL, et al. Effect of age on outcome of reduced-intensity hematopoietic cell transplantation for older patients with acute myeloid leukemia in first complete remission or with myelodysplastic syndrome. J Clin Oncol. 2010;28(11):1878-1887. [21] Büchner T, Berdel WE, Haferlach C, et al. Age-related risk profile and chemotherapy dose response in acute myeloid leukemia: a study by the German Acute Myeloid Leukemia Cooperative Group. J Clin Oncol. 2009;27(1):61-69. [22] Koreth J, Aldridge J, Kim HT, et al. Reduced-intensity conditioning hematopoietic stem cell transplantation in patients over 60 years: hematologic malignancy outcomes are not impaired in advanced age. Biol Blood Marrow Transplant. 2010;16(6):792-800. [23] Kim H, Kim BS, Kim DH, et al.Comparison between matched related and alternative donors of allogeneic hematopoietic stem cells transplanted into adult patients with acquired aplastic anemia: multivariate and propensity score-matched analysis. Biol Blood Marrow Transplant. 2011;17(9):1289-1298. [24] Kanda Y. Investigation of the freely available easy-to-use software ‘EZR’ for medical statistics. Bone Marrow Transplant. 2013;48(3):452-458. [ 25] Scott BL, Pasquini MC, Logan B, et al.Results of a Phase III Randomized, MultiCenter Study of Allogeneic Stem Cell Transplantation after High Versus Reduced Intensity Conditioning in Patients with Myelodysplastic Syndrome (MDS) or Acute Myeloid Leukemia (AML): Blood and Marrow Transplant Clinical Trials Network (BMT CTN) 0901. Blood. 2015;126(23): Abstract LBA-8.
Figure 1. Cumulative incidence of relapse (CIR) at 2 years using FBT 400 and FBT 200 in the overall population (A) and Propensity score matching (PSM) group (B).
16 Page 16 of 22
Table 1. Pre-transplant patient and disease characteristics
Overall population (pts, %)
age (years), median (range) Disease, no. (%) AML MDS HLA and donor type, no. (%) MRD MUD MMUD GVHD prophylaxis CSA + MMF TCD + CSA Others Cytogenetics Good/intermediate risk Adverse risk CR status CR1 CR2 MDS with blast < 10% HCT-CI, median (range)
248
PSM group FBT200 (n=121; 48.8%) 62 (20-71)
p value
84
FBT400 (n=42,50%)
FBT200 (n=4250%)
p value
54 (19-71)
FBT400 (n=127; 51.2%) 48 (18-61)
<0.001
56 (28-68)
55 (28-61)
58 (46-68)
0.009
191(77.0) 57(23.0)
102(80.3) 25(19.7)
89(73.5) 32(26.5)
0.206
68(81.0) 16(19.0)
32(76.1) 10(23.9)
36(85.7) 6(14.3)
0.266
103(41.5) 115(46.4) 30(12.1)
58(45.7) 50(39.3) 19(15.0)
45(37.2) 65(53.7) 11(9.1)
0.061
35(41.6) 38(45.3) 11(13.1)
19(45.2) 20(47.6) 3(7.1)
16(38.2) 18(42.8) 8(19.0)
0.268
152(61.3) 89(35.9) 7(2.8)
73(57.4) 50(39.3) 4(3.3)
79(65.3) 39(32.3) 3(2.4)
48(57.2) 34(40.5) 2(2.3)
23(54.7) 18(42.8) 1(2.5)
25(59.3) 16(38.2) 1(2.5)
0.904
194(78.3) 54(21.7)
103(81.1) 24(18.9)
91(74.3) 30(24.7)
61(72.6) 23(27.4)
32(76.2) 10(23.8)
29(69.0) 13(31.0)
132(53.0) 56(23.0) 60(24.0)
71(55.9) 34(26.8) 22(17.3)
67(55.4) 25(20.7) 29(23.9)
0.316
52(62.0) 15(17.8) 17(20.2)
25(59.5) 7(16.6) 10(23.9)
27(64.4) 8(19.0) 7(16.6)
0.714
1 (0-7)
1 (0 - 5)
2 (0 - 7)
<0.001
1.5 (0-7)
1 (0-5)
2 (0-7)
0.431
0.473
0.261
0.463
PSM: propensity score matching; pts: patients; FBT400: fludarabine (50 mg/m2/day for 4 days), busulfan (3.2 mg/kg/day for 4 days) and total body irradiation (TBI) 400 cGy; FBT200: fludarabine (30mg/m2/day for 4 days), busulfan (3.2 mg/kg/day for 2 days), and total body irradiation (TBI) 200 cGy; AML: Acute myeloid Leukemia; MDS: myelodysplastic syndrome; MRD: matched related donor; MUD: matched unrelated donor; MMUD: mismatched unrelated donor; GVHD: graft-versus-host disease; CSA:
cyclosporine; MMF: mycophenolate
mofetil; TCD: in vivo T-cell depletion; CR: complete remission; HCT-CI: hematopoietic cell transplantationcomorbidity index.
17 Page 17 of 22
Table 2. Comparison of outcomes between FBT 400 and FBT200.
Overall population Acute GVHD grade 1-4 grade 2-4 grade 3/4 Chronic GVHD at 1 y (any grade) Outcomes at 2 y OS NRM Relapse
PSM group
Overall % ± SE
FBT400 (n=127)
FBT200 (n=121)
p value
Overall % ± SE
FBT400 (n=42)
FBT200 (n=42)
p value
59.8±5.9 43.6±4.3 19.4±1.9
63.8±6.3 46.5±4.6 21.3±2.1
55.6±5.5 40.6±4.0 17.4±1.7
0.171 0.086 0.479
61.9±6.0 41.7±4.1 16.7±1.7
71.4±6.8 47.6±4.6 21.3±2.1
52.4±5.1 35.7±3.5 17.4±1.7
0.143 0.091 0.479
53.9±5.3
51.4±5.0
56.5±5.5
0.605
52.1±5.1
39.3±3.9
64.8±6.2
0.138
48.0±3.6 34.6±3.6 24.8±3.5
51.7±5.2 34.1±4.6 14.1±1.4
45.2±5.0 28.7±2.8 26.1±2.6
0.541 0.368 0.033
48.0±3.6 34.6±3.6 24.8±3.5
54.7±7.7 33.3±2.8 9.5±1.1
49.5±9.1 22.2±2.3 25.7±2.6
0.718 0.238 0.315
PSM: propensity score matching; GVHD: graft-versus-host disease; y: year; OS: overall survival; NRM: nonrelapse mortality.
18 Page 18 of 22
Table 3. UVA and MVA of 2-year OS in overall population and PSM group
Overall population UVA Parameter Age (years) Disease state at HSCT Donor type
Diagnosis Conditioning Cytogenetics HCT-CI Period effect
CR1 CR2 No CR MRD MUD MMUD AML MDS FBT400 FBT200 Fav/Interm Adverse ordinal continuous
MVA
UVA
PSM group MVA
HR [95% CI]
p value
HR [95% CI]
p value
HR [95% CI]
p value
HR [95% CI]
1.023 [1.005-1.042] 1.000 1.834 [1.096-2.816]
0.014 0.007
1.031 [1.011-1.051] 1.000 1.424 [1.0960-1.714]
0.001 0.018
1.012 [0.961-1.065] 1.000 1.013 [0.404-2.540]
0.640 0.978
1.012 [0.950-1.078] 1.000 1.300 [0.493-3.428]
0.710 0.595
1.000 2.500 [1.490-4.201]
0.002
-
-
1.000 1.508 [0.719-3.161]
0.002
-
1.000 1.628 [1.094-2.423] 1.000 1.117 [1.091-3.703] 1.000 1.628 [1.091-2.427] 1.124 [1.009-1.252] 1.020 [0.921-1.131]
0.015
1.000 1.464 [0.920-2.331] 1.000 1.606 [1.054-2.448] -
-
1.000 2.659 [1.328-5.319] 1.000 1.124 [0.594-2.132] 1.000 3.485 [1.796-6.761] 0.998 [0.829-1.202] 1.134 [0.931-1.381]
0.006
1.000 0.670[0.3427 -1.311] 1.000 2.574 [1.467-5.969] -
0.541 0.016 0.155 0.243
0.107 0.027 -
0.718 <0.001 0.980 0.211
p value
0.240 0.002 -
PSM: propensity score matching; UVA: univariate analysis; MVA: multivariate analysis; HR: hazard ratio; HSCT: hematopoietic stem cell transplant (allogeneic); CR: complete remission; MRD: matched related donor; MUD: matched unrelated donor; MMUD: mismatched unrelated donor; AML: Acute myeloid Leukemia; MDS: myelodysplastic syndrome; FBT400: fludarabine (50 mg/m2/day for 4 days), busulfan (3.2 mg/kg/day for 4 days) and total body irradiation (TBI) 400 cGy; FBT200: fludarabine (30mg/m2/day for 4 days), busulfan (3.2 mg/kg/day for 2 days), and total body irradiation (TBI) 200 cGy; Fav/interm: favorable/intermediate; HCT-CI: hematopoietic cell transplantation-comorbidity index.
19 Page 19 of 22
Table 4. UVA and MVA of NRM in Overall population and PSM group
Overall population UVA Parameter Age (years) Disease state at HCT Donor type
Diagnosis Conditioning Cytogenetics HCT-CI Period effect
HR [95% CI] CR1 CR2 No CR MRD MUD MMUD AML MDS FBT400 FBT200 Fav/Interm Adverse ordinal continuous
p value
1.012 [0.990-1.033] 1.000 1.465 [1.125-1.906]
0.280 0.004
1.000 1.79[1.320-2.427]
0.001
1.000 2.183 [1.353-3.521] 1.000 0.855[0.544-1.344] 1.000 1.385[0.809-2.372] 1.117[0.971-1.284] 1.064[0.931-1.215]
0.001 0.500 0.230 0.120 0.360
MVA HR [95% CI] 1.000 1.759 [1.298-2.384] 1.000 2.044 [1.256-3.322] 1.000 0.806 [0.495-1.314] -
UVA p value
HR [95% CI]
PSM group MVA p value
HR [95% CI]
p value
-
0.992 [0.926-1.064] 1.000 1.919 [1.186-3.104]
0.840 0.008
1.000 1.317 [0.547-3.166]
0.540
0.002
1.000 2.195 [1.302-3.701]
0.003
1.000 2.144 [1.085-4.236]
0.028
0.004
1.000 0.306 [0.133-0.701] 1.000 0.609 [0.267-1.390] 1.000 2.570 [1.132-5.839] 0.749 [0.544-1.031] 1.266 [0.989-1.621]
0.005
1.000 0.650 [0.125-3.378] 1.000 0.516 [0.206-1.293] -
0.610
0.390 -
0.240 0.024 0.077 0.061
0.160 -
PSM: propensity score matching; UVA: univariate analysis; MVA: multivariate analysis; HR: hazard ratio; HSCT: hematopoietic stem cell transplant (allogeneic); CR: complete remission; MRD: matched related donor; MUD: matched unrelated donor; MMUD: mismatched unrelated donor; AML: Acute myeloid Leukemia; MDS: myelodysplastic syndrome; FBT400: fludarabine (50 mg/m2/day for 4 days), busulfan (3.2 mg/kg/day for 4 days) and total body irradiation (TBI) 400 cGy; FBT200: fludarabine (30mg/m2/day for 4 days), busulfan (3.2 mg/kg/day for 2 days), and total body irradiation (TBI) 200 cGy; Fav/interm: favorable/intermediate; HCT-CI: hematopoietic cell transplantation-comorbidity index.
20 Page 20 of 22
Table 5. UVA and MVA of CIR in Overall population and PSM group.
Overall population UVA Parameter Age (years) Disease state at HCT
HR [95% CI]
p value
MVA HR [95% CI]
UVA p value
HR [95% CI]
PSM group MVA p value
HR [95% CI]
CR1 CR2
1.039 [1.006-1.073] 1.000 1.035 [0.755-1.416]
0.021 0.830
1. 045 [1.011-1.081] 1.000 1.833 [1.098-3.061]
0.009 0.020
1.039 [0.969-1.113] 1.000 1.116 [0.625-1.992]
0.280 0.710
1.005 [0.918-1.100] 1.000 1.190 [0.429-3.303]
Donor type
No CR MRD MUD
1.000 0.834 [0.531-1.311]
0.43
-
-
1.000 1.089 [0.478-2.518]
0.850
-
Diagnosis
MMUD AML MDS
1.000 1.437 [0.698-2.954]
0.353
0.006
1.000 1.802 [0.429-7.571]
0.42
FBT400 FBT200 Fav/Interm Adverse ordinal continuous
1.000 1.932 [1.087-3.433] 1.000 1.675 [0.917-3.057] 1.437 [0.6986-2.954] 0.9232 [0.7962-1.07]
0.025
1.000 4.148 [1.49411.520] 1.000 1.341 [0.662-2.714] 1.000 2.053 [1.083-3.890] -
0.410
1.000 1.683 [0.598-4.729] 1.000 2.831 [1.047-7.657] 1.387 [1.131-1.700] 0.793 [0.588-1.071]
0.32
Conditioning Cytogenetics HCT-CI Period effect
0.093 0.320 0.290
0.027 -
0.04 0.001 0.130
1.000 4.236 [0.627-28.610] 1.000 1.428 [0.448-4.552] 1.000 3.813 [1.162-12.510] -
p value 0.910 0.740
0.140
0.550 0.027 -
PSM: propensity score matching; UVA: univariate analysis; MVA: multivariate analysis; HR: hazard ratio; HSCT: hematopoietic stem cell transplant (allogeneic); CR: complete remission; MRD: matched related donor; MUD: matched unrelated donor; MMUD: mismatched unrelated donor; AML: Acute myeloid Leukemia; MDS: myelodysplastic syndrome; FBT400: fludarabine (50 mg/m2/day for 4 days), busulfan (3.2 mg/kg/day for 4 days) and total body irradiation (TBI) 400 cGy; FBT200: fludarabine (30mg/m2/day for 4 days), busulfan (3.2 mg/kg/day for 2 days), and total body irradiation (TBI) 200 cGy; Fav/interm: favorable/intermediate; HCT-CI: hematopoietic cell transplantation-comorbidity index.
21 Page 21 of 22
Figure 1.
22 Page 22 of 22
S1083-8791(16)30327-5 http://dx.doi.org/doi: 10.1016/j.bbmt.2016.08.030 YBBMT 54385
To appear in:
Biology of Blood and Marrow Transplantation
Received date: Accepted date:
7-7-2016 31-8-2016
Please cite this article as: Hassan Sibai, Umberto Falcone, Uday Deotare, Fotios V. Michelis, Jieun Uhm, Vikas Gupta, John Kuruvilla, Jeffrey H. Lipton, Matthew D. Seftel, Hans A. Messner, Dennis (Dong Hwan) Kim, Myeloablative Versus Reduced-Intensity Conditioning in Patients with Myeloid Malignancies: a Propensity Score Matched Analysis., Biology of Blood and Marrow Transplantation (2016), http://dx.doi.org/doi: 10.1016/j.bbmt.2016.08.030. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Myeloablative versus reduced-intensity conditioning in patients with myeloid malignancies: a propensity score matched analysis. Hassan Sibai*, MD, Umberto Falcone*, MD, PhD, Uday Deotare, MD, Fotios V. Michelis, MD, PhD, Jieun Uhm, MD, PhD, Vikas Gupta, MD, John Kuruvilla, MD, Jeffrey H. Lipton, MD, Matthew D. Seftel, MD, Hans A. Messner, MD, and Dennis (Dong Hwan) Kim, MD, PhD. Allogeneic Blood and Marrow Transplant Program, Princess Margaret Cancer Centre, University Health Network, University of Toronto, Toronto, Canada. *Both these authors contributed equally to this work.
Corresponding author: Dennis (Dong Hwan) Kim, MD, PhD Rm 5-126, Department of Medical Oncology and Hematology Princess Margaret Cancer Center, 610, University Ave., Toronto, Canada M5G 2M9 Tel: 1-416-946-4501 x 2464 Fax: 1-416-946-4563 Email: [email protected] Presented in part at the 56th American Society of Hematology Annual Meeting in San Francisco, California, December 6-9, 2014.
Short title: Propensity score analysis of allo-HSCT conditioning regimens.
Financial Disclosure Statement: The authors state that they have no conflict of interests
1 Page 1 of 22
HIGHLIGHTS RIC is comparable to MAC for AML/MDS PSM analysis balances pretransplant patient characteristics Addition of low-dose TBI to allo-SCT conditioning could reduce the risk of relapse
Abstract Reduced-intensity conditioning (RIC) has been shown to have similar overall survival (OS) but higher relapse rates when compared to myeloablative (MAC) regimens in patients with myeloid malignancies. Using propensity score matching (PSM) analysis, well-balanced pairs of different variables can be compared effectively. We retrospectively compared allogeneic hematopoietic stem cell transplantation patients with acute myeloid leukemia and myelodysplasia receiving RIC regimen (FBT 200; fludarabine 30mg/m2/day for 4 days, busulfan 3.2 mg/kg/day for 2 days, and total body irradiation (TBI) 200 cGy) or MAC regimen (FBT 400; fludarabine 50 mg/m2/day for 4 days, busulfan 3.2 mg/kg/day for 4 days and TBI 400 cGy). Two hundred and forty-eight patients (127 MAC, 121 RIC) were included in the analysis. No statistical difference was observed in 2-year OS (RIC 45.2±5.0%, MAC 51.7±5.2%; p=0.541), non-relapse mortality (NRM, RIC 28.7±2.8% MAC 34.7±4.6% p=0.368), and acute (p=0.171) or chronic (p=0.605) GVHD at one year between FBT200 and FBT400. Cumulative incidence of relapse at 2 years (CIR) was statistically different (RIC 26.1±2.6%, MAC 14.2±3.5%; p=0.033). When PSM was applied to the study population, 42 case-control pairs were evenly matched. PSM analysis confirmed no statistical difference in 2-year OS (RIC 49.0±9.1%, MAC 54.9±7.7%; p=0.718), NRM (RIC 22.2 ± 2.3%, MAC 33.3±2.8%;
2 Page 2 of 22
p=0.238). CIR did not differ statistically between FBT200 and FBT400 (RIC 25.7±2.6%, MAC 9.5±1.1%; p=0.315) in the PSM pairs. Our study shows that after applying PSM, FBT 200 RIC conditioning is able to give OS, NRM, and CIR rates comparable to FBT 400 MAC conditioning.
Keywords: Propensity score matching; allogeneic stem cell transplant; conditioning regimen; acute myeloid leukemia; myelodysplasia.
3 Page 3 of 22
Introduction Allogeneic hematopoietic stem cell transplantation (allo-HSCT) is a curative treatment
modality for patients with acute myeloid
leukemia (AML)
and
myelodysplastic syndrome (MDS) as well as other hematological disorders. 1-4 Myeloablative conditioning (MAC) regimens, such as high doses of total body irradiation (TBI) plus chemotherapy2 or combination chemotherapy regimens including busulfan and cyclophosphamide1 have been widely used while the use of reduced intensity conditioning (RIC) has been increasing over the past 2 decades.7-9 RIC includes reduced dose of chemotherapy and/or low dose TBI. Adding low dose TBI to RIC regimens has been shown to be a well-tolerated option with anti-leukemic activity comparable to MAC regimens.17 However it is still in debate whether RIC results in similar post-transplant outcomes compared to MAC. While suspected to increase the risk of relapse due to lower intensity of chemotherapy, RIC is associated with less conditioning-related toxicities and non-relapse mortality (NRM) when compared to MAC.5-7, 10-11 In addition to recipient age, other factors including disease status at transplant, donor type, cytogenetics, hematopoietic cell transplantation-comorbidity index (HCT-CI)
12
,
contribute to the outcome variables in both MAC and RIC transplants. In most transplant studies, these pre-transplant factors are not entirely matched between comparable groups. The propensity score is the probability of treatment assignment conditional on observed baseline characteristics. Propensity score matching (PSM) analysis is used to balance the variables affecting the choice of a treatment among different treatment options.13, 14 We conducted a retrospective study including patients undergoing allo-HST for AML or MDS, with the aim to compare MAC and RIC conditioning regimens, using PSM
4 Page 4 of 22
analysis to obtain well matched pairs of patients. Our hypothesis is that by this mean, OS and relapse rate obtained with RIC are comparable to those obtained with MAC.
Patients and Methods Study design and eligibility criteria We retrospectively compared MAC and RIC conditioning regimens in patients who underwent allo-HST for AML or MDS at Princess Margaret Cancer Centre (PMCC) between January 2009 and December 2013. This study was approved by the Research Ethics Board of the University Health Network/Princess Margaret Cancer Centre. Only AML patients in first (CR1) or second complete remission (CR2) and MDS patients with <10% bone marrow blasts at the time of the transplant were included in the study. For AML, cytogenetic risk at diagnosis was characterized as Favorable/intermediate versus Adverse, as described by SWOG/ECOG.15 In patients with good cytogenetic risk allo-HST was offered for AML refractory to first-line treatment, or at relapse after achieving CR2. For MDS cytogenetics, 2008 WHO MDS cytogenetic minimal criteria were used.16 Inclusion was restricted to patients receiving Fludarabine/Busulfan (FB) plus TBI with either RIC (Fludarabine 30mg/m2/day for 4 days, Busulfan 3.2mg/kg/day for 2 days and TBI 200 cGy) or MAC conditioning (Fludarabine 50mg/m2/day for 4 days, Busulfan 3.2mg/kg/day for 4 days and TBI 400 cGy). The decision to offer RIC was primarily based on patient’s age (≥60 years) and/or the presence of comorbidities.5, 20-22
Stem cell source consisted of matched-related (MRD), matched-unrelated
(MUD), or mismatched unrelated (MMUD) donors. Donor peripheral blood stem cells (PBSC) were mobilized with G-CSF. Haploidentical or cord blood allograft recipients were excluded from this study. Graft-versus-host disease (GVHD) prophylaxis in 5 Page 5 of 22
MRD allo-HSCT consisted of cyclosporine A (CSA) combined with either mycophenolate mofetil (MMF, 15 mg/kg by mouth or IV twice daily (dose was rounded to the nearest multiple of 250) from day 0 for 30 days and then stopped without taper) or methotrexate (15 mg/m2 on HSCT day +1 and 10 mg/kg on HSCT days +3, +6, and +11). In patients receiving MUD allo-HSCT, T-cell depletion was performed using low-dose alemtuzumab or rabbit antithymocyte globulin (ATG) in combination with CSA. Pre-transplant variables included in the PSM were: age, HCT-CI, remission status (CR1 vs CR2, and MDS with <10 % blasts) at the time of transplant, diagnosis (AML vs MDS), cytogenetic risk group (adverse-risk vs favorable (relapsed) and intermediate risk), donor type (MRD vs MUD and MMUD), and period effect (year of transplant). We adopted PSM analysis to adjust the risk factors affecting the choice of conditioning regimens by doing well balanced pairs of RIC and MIC patients.
Definitions of statistical endpoints Overall survival (OS) duration was measured from the date of allo-HSCT until death from any causes. Alive patients were censored on the date of their last follow-up. Cumulative incidence of relapse (CIR) was calculated from the date of allo-HSCT until relapse. Relapse was defined as ≥5% blasts in a bone marrow aspirate or peripheral blood, or the development of extra-medullary leukemia following alloHSCT. Non-relapse mortality (NRM) was calculated as death without evidence of disease relapse. Acute and chronic GVHD were diagnosed and graded using established criteria. 18-19
Statistical analysis
6 Page 6 of 22
Patient and disease characteristics were reported using descriptive statistics. Data were updated as of December 2013.The main outcome variables of interest included: OS, cumulative incidence of relapse (CIR) and cumulative incidence of NRM. CIR and NRM are competing risks setting. For relapse incidence, death from non-relapse cause was accounted as a competing event, while relapse was taken account for competing event for NRM. Cumulative incidence of grade II-IV acute GVHD and chronic GVHD at one year were estimated taking into account death as a competing event. In order to adjust for any potential biases derived from imbalanced pretranplant factors between FBT200 vs FBT400 groups, we have adopted propensity score matching analysis. For an initial step, propensity score was calculated initially using binary logistic regression model. The following 7 independent pre-transplant factors were included into the binary logistic regression model for calculation of propensity score: age, remission status, diagnosis, cytogenetic risk group (adverserisk vs favorable/intermediate), donor type, HCT-CI and period effect. Then each patient from FBT200 group was matched and paired to another patient from FBT400 treatment group with the lowest differences of propensity score using in-house script. For risk factor analysis, we have included the 7 covariates previously described with conditioning regimen (i.e. MAC vs RIC) in the analysis for OS, CIR and NRM. These analyses were performed repetitively in the whole cohort of patients and in the PSM groups, separately. After the univariable analysis, multivariable analysis adopted Cox’s hazard proportional regression model for OS and Fine-Grey method for CIR and NRM. Stepwise selection procedure was applied using the criteria for variable selection, p=0.05 for variable entry and p=0.1 for variable removal. In addition to all the variables remained in the multivariable model, the variable of interest which is conditioning regimen is included in the final model as shown in Table 3 and 4. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated for the significant risk factors. Paired analysis was adopted throughout the PSM analysis for survival. For MVA in PSM groups, only the variables statistically significant (plus 7 Page 7 of 22
conditioning regimen) at MVA for the whole cohort were considered. Statistical analyses were performed using EZR on R Commander Version 1.11.24
Results Patients, diseases and treatment characteristics Patient characteristics are summarized in Table 1. A total of 248 patients (male: 134; 54%) were transplanted using PBSC as a graft source. Median age at transplant for the entire cohort was 54 years (range: 18-71). FBT400 was administered to 127 patients (51.2%) and FBT200 to 121 (48.8%). Remission status was CR1 in 132 patients (53%) while 56 patients (23%) underwent transplant in CR2. Adverse risk cytogenetics was present in 54 patients (21.7%) while patients with good and intermediate risk cytogenetics were 194 (78.3%). Source of stem cells was MRD in 103 cases (41.5%), while MUD and MMUD were 115 (46.4%) and 30 (12.1%), respectively. Demographics and pre-transplant characteristics were imbalanced between FBT200 and FBT400 groups. Older age (P<0.001), higher HCT-CI score (p=0.001) and more MRD (p=0.061) were observed in the RIC group. However, no differences were observed between FBT200 and FBT400 groups in CR status at HSCT (p=0.316), subtype of diagnosis (AML vs MDS, p=0.206), or cytogenetic risk group (p=0.261). Using PSM analysis to overcome baseline imbalances, 42 case-control pairs (84 patients) were selected (Table 1). Pre-transplant variables became well balanced between the two groups (FBT400 and FBT200) after applying PSM. Median age at transplant remained different between the 2 PSM groups (55 years in FBT400, 58 years in FBT200; p=0.009).
8 Page 8 of 22
Overall outcomes With a median follow-up of 18 months among survivors in the overall population (n=248), the 2-year OS, NRM and relapse incidence rate was 48.0±3.6%, 34.6±3.6% and 24.8±3.5%, respectively. There was no difference in OS (45.2±5.0% in RIC vs 51.7±5.2% in MAC patients at 2 years; p=0.541) or NRM (28.7±2.8% in RIC vs 34.7±4.6% in MAC patients at 2 years; p=0.368) between FBT200 and FBT400 groups. The CIR at 2 years was higher in the RIC group (26.1±2.6% in RIC vs 14.1±1.4% in MAC; p=0.033, Figure 1A). In the PSM subgroup of patients, no statistical difference was noted in 2-year OS (RIC 49.0±9.1%, MAC 54.9±7.7%; p=0.718), NRM (RIC 22.2±2.3%, MAC 33.3±2.8%; p=0.238), or CIR (RIC 25.7±2.6%, MAC 9.5±1.1%; p=0.315, Figure 1B) between FBT200 and FBT400 groups.
Acute and chronic GVHD according to conditioning for overall and PSM pairs The overall incidence of grade 3-4 acute GVHD was 19.4±1.9%. When comparing between the two conditioning regimens, FBT400 and FBT200, there was no difference in rates of acute and chronic GVHD of any grade (Table 2). PSM pairs showed no difference for acute (any grade; FBT400 71.4% vs FBT200 52.4%, p=0.143) or chronic (any grade; FBT400 39.3% vs FBT200 64.8%, p=0.138) GVHD rate at one year (Table 2).
Univariate and Multivariate analysis of prognostic factor for transplant outcomes Overall Survival
9 Page 9 of 22
In the whole population, age ≤60 years (p=0.014), CR1 (p= 0.007), diagnosis of AML (p= 0.015), related donor (p=0.002) and favorable/intermediate risk cytogenetics (p= 0.016) were all predictors of better OS at UVA. Only age (p=0.01), disease state at HSCT (p=0.018) and cytogenetic risk group (0.027) remained significant at multivariate analysis (MVA). Conditioning regimen was statistically not significant at UVA (p=0.541). In PSM pairs, UVA for the same variables (Table 3) showed no differences in OS for the 2 conditioning regimens (all patients: 48.0±3.6 months, FBT400: 54.7±7.7 months, FBT200 49.5±9.1 months; p=0.718). Statistically significant variables were donor type (p=0.002), diagnosis (p=0.006) and cytogenetic risk group (p<0.001). At MVA, only cytogenetic risk group remained statistically significant, with adverse risk group
showing
a
worse
OS
(15.6±9.1
months
vs
66.4±6.4
months
in
favorable/intermediate risk; p=0.002).
Non-relapse mortality In the overall population of patients, disease status (p=0.004), donor type (p=0.001), and diagnosis (p=0.001) were significant at UVA. No impact of conditioning regimen was observed on NRM even when PSM was applied. At MVA only donor type (p=0.002) and diagnosis (p=0.004) remained significant (Table 4). In PSM pairs, CR2 disease state at transplant (p=0.008), MMUD type (p=0.003), AML diagnosis (p=0.005), and adverse cytogenetic risk group (p=0.024) were related to higher NRM at 2 years.
Transplant from MRD (vs MUD/MMUD) resulted in
improved NRM at MVA {HR=1.031; 95% CI [1.085-4.236]; p= 0.028}.
10 Page 10 of 22
Cumulative incidence of relapse UVA for the whole patient cohort showed that CIR at 2 years was adversely affected by age >60 years (p=0.021) and RIC regimen (vs MAC; HR=1.932; 95% CI [1.0873.433]; P=0.033). At MVA, CIR was higher in patients aged >60 years (p=0.009), not in CR1 (p=0.020), with MDS diagnosis (p=0.006), and adverse risk cytogenetics (p=0.027). Conditioning regimen did not remain statistically significant at MVA {FBT200: HR=1.341, 95% CI [0.662-2.714]; p=0.401} (Table 5). For CIR in PSM pairs, only adverse cytogenetic risk group remained significant at both UVA and MVA (Table 5).
Discussion A previous study at a single institution comparing a MAC regimen including IV busulfan plus cyclophosphamide (ivBUCY) with two RIC regimens of fludarabine plus IV busulfan 6.4 mg/kg (FB2) or IV busulfan 12.8 mg/kg (FB4) in patients with AML and MDS at various stages of disease, showed no difference in 2-year OS (ivBUCY 50%, FB2 47%, FB4 49%; p=NS).5 Kharfan-Dabaja et al. reviewed FB2 and FB4 regimens in patients with AML in CR1 undergoing allo-HSCT at various participating centers from European Bone Marrow Transplant Group (EBMT) showing that FB2 can be safely used in these patients when aged ≥ 50 years while it does not affect OS when used in patients <50 years.6
11 Page 11 of 22
Recently, randomized BMT CTN 0901 study was stopped prematurely because RIC resulted in higher relapse rates and lower transplant related mortality (TRM) compared to MAC, with a statistically significant advantage in RFS for patients receiving MAC; however the right choice of conditioning intensity remains challenging with a necessity to balance the risk of relapse with TRM.25 In our study of 248 patients with AML or MDS, we found that with a median follow-up of 18 months there was a higher incidence of relapse in the RIC group. However, since there was a fundamental difference of age criteria between the 2 groups, with MAC used for patients ≤60 years and RIC for pts >60 years, this analysis was further subjected to PSM. After balancing the variables, we found 42 case/control pairs between the 2 groups. We were able to conclude that there were no statistical differences in OS, NRM, or relapse rate. An interesting finding was that chronic GVHD rates at 1 year were 64.8% vs 39.3% in RIC and MAC arms, respectively. Although statistical significance was not reached, a trend towards higher rates of chronic GVHD in RIC patients was observed. Since the incidence of chronic GVHD is associated with reduced relapse rates, we postulated that unevenness in the initial findings was effectively highlighted in the PSM analysis. PSM analysis has not been frequently used as a statistical method for comparison of groups in allo-HSCT studies. PSM is able to match pairs in an efficient manner across groups. While previously PSM has been used in this setting for patients with acquired aplastic anemia undergoing allo-HSCT 23, to the best of our knowledge, this study is the first to use PSM to compare conditioning regimens (RIC vs MAC) for alloHCT in hematological malignancies, in a sizable number of patients. It is realistically impossible to design a prospective clinical trial randomizing MRD vs alternative donor for allo-HSCT in aplastic anemia. However, by adopting PSM, they
12 Page 12 of 22
have used the score by comparing 25 patients in each group of MRD, MUD, and MMUD transplants from a total of 225 acquired aplastic anemia patients, thus balancing imbalanced variables using PSM.23 In our study, negation of significant variables was found so that PSM is needed as an equalizer in multiple variables in allo-HSCT studies and hence deserves due recognition in the current world. In our study we were able to group-match age, disease risk status, donor type, diagnosis, conditioning regimen, cytogenetic risk group, comorbidity score (HCT-CI), and period of transplant. When we compared our results of FB plus low dose TBI to previous experiences with FB without radiation 6, we found that the addition of low-dose TBI was able to reduce the relapse rate at 2 years: 9.5% our study (PSM), 23% Kharfan-Dabaja et al) although it is indirect comparison. This was seen with both MAC and RIC regimens, although the relapse rates were higher in RIC as compared to MAC. Previously, Russell et al.17 showed that the addition of low dose TBI (400cGy) to the FB regimen was able to reduce the relapse risk (HR= 0.29) as well as the remaining parameters, including OS and DFS. Hence our data suggest that the addition of low-dose TBI to FB combination could reduce the risk of relapse. In addition to the inherent limitations of the retrospective nature of the study, the patient comparable group was small. By applying PSM the overall sample size came down from 248 to 84. The pre-transplant mutational status (i.e. nucleophosmin-1 (NPM1) mutation, FMS-related tyrosine kinase-3 (FLT-3) internal tandem duplication mutation), chemotherapy treatment for the myeloid malignancy, GVHD prophylaxis regimen, and CMV status at the time of transplant were not taken into consideration at the time of analysis. In conclusion, our study shows that RIC conditioning using combination of FB+TBI (200cGy) is able to give similar survival rates and CIR rates compared to FBT400
13 Page 13 of 22
MAC conditioning. Outside of the propensity model, the CIR at 2 years was higher in the RIC group. This conditioning regimen will require further studies and evaluation in the years to come.
Acknowledgements: Previously presented in part at the 56th American Society of Hematology Annual Meeting in San Francisco, California, December 6-9, 2014. Conflict of Interest: The authors declare no conflict of interest. Authorship Contributions: HS and UF contributed equally to this work. HS, UF, DK, UD analyzed and interpreted data and drafted the paper. HS, JU, and FM collected the data. HS, DK, UF, UD: Concept and designed the study. HS, DK performed the statistical analysis. HS, DK, UF, UD, FM, JU, VK, JK, JL, MS, and HM critically revised the manuscript.
References [1] Santos GW, Tutschka PJ, Brookmeyer R, Saral R, Beschorner WE, Bias WB et al. Marrow transplantation for acute nonlymphocytic leukemia after treatment with busulfan and cyclophosphamide. N Engl J Med. 1983;309(22):1347-1353. [2] Blume KG, Beutler E, Bross KJ, Chillar RK, Ellington OB, Fahey JL et al. Bonemarrow ablation and allogeneic marrow transplantation in acute leukemia. N Engl J Med. 1980;302(19):1041-1046. [3] Deeg HJ. Hematopoietic cell transplantation for myelodysplastic syndrome. Am Soc Clin Oncol Educ Book. 2015:e375-80. [4] Brown JR, Kim HT, Armand P, Cutler C, Fisher DC, Ho V et al. Long-term followup of reduced-intensity allogeneic stem cell transplantation for chronic lymphocytic leukemia: prognostic model to predict outcome. Leukemia. 2013;27(2):362-9. [5] Shimoni A, Hardan I, Shem-Tov N, Yeshurun M, Yerushalmi R, Avigdor A et al. Allogeneic hematopoietic stem-cell transplantation in AML and MDS using myeloablative versus reduced-intensity conditioning: the role of dose intensity. Leukemia. 2006;20(2):322-328. 14 Page 14 of 22
[6] Kharfan-Dabaja MA, Labopin M, Bazarbachi A, et al. Comparing i.v. BU dose intensity between two regimens (FB2 vs FB4) for allogeneic HCT for AML in CR1: a report from the Acute Leukemia Working Party of EBMT. Bone Marrow Transplant. 2014;49(9):1170-1175. [7] Slavin S, Nagler A, Naparstek E, Kapelushnik Y, Aker M, Cividalli G et al. Nonmyeloablative stem cell transplantation and cell therapy as an alternative to conventional bone marrow transplantation with lethal cytoreduction for the treatment of malignant and nonmalignant hematologic diseases. Blood. 1998;91(3):756-63. [8] Socie G, Clift RA, Blaise D, Devergie A, Ringden O, Martin PJ et al. Busulfan plus cyclophosphamide compared with total-body irradiation plus cyclophosphamide before marrow transplantation for myeloid leukemia: long-term follow-up of 4 randomized studies. Blood. 2001;98(13):3569-3574. [9] Reshef R, Porter DL. Reduced-intensity conditioned allogeneic SCT in adults with AML. Bone Marrow Transplant. 2015;50(6):759-769. [10] Alyea EP, Kim HT, Ho V, Cutler C, Gribben J, DeAngelo DJ et al. Comparative outcome of nonmyeloablative and myeloablative allogeneic hematopoietic cell transplantation for patients older than 50 years of age. Blood. 2005;105(4):18101814. [11] Gooley TA, Chien JW, Pergam SA, Hingorani S, Sorror ML, Boeckh M et al. Reduced mortality after allogeneic hematopoietic-cell transplantation. N Engl J Med. 2010;363(22):2091-2101. [12] Sorror ML, Maris MB, Storb R, Baron F, Sandmaier BM, Maloney DG et al. Hematopoietic cell transplantation (HCT)-specific comorbidity index: a new tool for risk assessment before allogeneic HCT. Blood. 2005;106(8):2912-2919. [13] Austin PC. An Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Studies. Multivariate Behav Res. 2011;46(3):399424. [14] Rosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects. Biometrika. 1983;70(1):41-55. [15] Slovak ML, Kopecky KJ, Cassileth PA, et al. Karyotypic analysis predicts outcome of preremission and postremission therapy in adult acute myeloid leukemia: a Southwest Oncology Group/Eastern Cooperative Oncology Group study. Blood. 2000;96(13):4075-4083. [16] Vardiman JW, Thiele J, Arber DA, et al. The 2008 revision of the World Health Organization (WHO) classification of myeloid neoplasms and acute leukemia: rationale and important changes. Blood. 2009;114(5):937-951. [17] Russell JA, Irish W, Balogh A, et al. The addition of 400 cGY total body irradiation to a regimen incorporating once-daily intravenous busulfan, fludarabine, and
15 Page 15 of 22
antithymocyte globulin reduces relapse without affecting nonrelapse mortality in acute myelogenous leukemia. Biol Blood Marrow Transplant. 2010;16(4):509-514. [18] Przepiorka D, Weisdorf D, Martin P, et al. 1994 Consensus Conference on Acute GVHD Grading. Bone Marrow Transplant. 1995;15(6):825-828. [19] Pavletic SZ, Lee SJ, Socie G, Vogelsang G. Chronic graft-versus-host disease: implications of the National Institutes of Health consensus development project on criteria for clinical trials. Bone Marrow Transplant. 2006;38(10):645-651. [20] McClune BL, Weisdorf DJ, Pedersen TL, et al. Effect of age on outcome of reduced-intensity hematopoietic cell transplantation for older patients with acute myeloid leukemia in first complete remission or with myelodysplastic syndrome. J Clin Oncol. 2010;28(11):1878-1887. [21] Büchner T, Berdel WE, Haferlach C, et al. Age-related risk profile and chemotherapy dose response in acute myeloid leukemia: a study by the German Acute Myeloid Leukemia Cooperative Group. J Clin Oncol. 2009;27(1):61-69. [22] Koreth J, Aldridge J, Kim HT, et al. Reduced-intensity conditioning hematopoietic stem cell transplantation in patients over 60 years: hematologic malignancy outcomes are not impaired in advanced age. Biol Blood Marrow Transplant. 2010;16(6):792-800. [23] Kim H, Kim BS, Kim DH, et al.Comparison between matched related and alternative donors of allogeneic hematopoietic stem cells transplanted into adult patients with acquired aplastic anemia: multivariate and propensity score-matched analysis. Biol Blood Marrow Transplant. 2011;17(9):1289-1298. [24] Kanda Y. Investigation of the freely available easy-to-use software ‘EZR’ for medical statistics. Bone Marrow Transplant. 2013;48(3):452-458. [ 25] Scott BL, Pasquini MC, Logan B, et al.Results of a Phase III Randomized, MultiCenter Study of Allogeneic Stem Cell Transplantation after High Versus Reduced Intensity Conditioning in Patients with Myelodysplastic Syndrome (MDS) or Acute Myeloid Leukemia (AML): Blood and Marrow Transplant Clinical Trials Network (BMT CTN) 0901. Blood. 2015;126(23): Abstract LBA-8.
Figure 1. Cumulative incidence of relapse (CIR) at 2 years using FBT 400 and FBT 200 in the overall population (A) and Propensity score matching (PSM) group (B).
16 Page 16 of 22
Table 1. Pre-transplant patient and disease characteristics
Overall population (pts, %)
age (years), median (range) Disease, no. (%) AML MDS HLA and donor type, no. (%) MRD MUD MMUD GVHD prophylaxis CSA + MMF TCD + CSA Others Cytogenetics Good/intermediate risk Adverse risk CR status CR1 CR2 MDS with blast < 10% HCT-CI, median (range)
248
PSM group FBT200 (n=121; 48.8%) 62 (20-71)
p value
84
FBT400 (n=42,50%)
FBT200 (n=4250%)
p value
54 (19-71)
FBT400 (n=127; 51.2%) 48 (18-61)
<0.001
56 (28-68)
55 (28-61)
58 (46-68)
0.009
191(77.0) 57(23.0)
102(80.3) 25(19.7)
89(73.5) 32(26.5)
0.206
68(81.0) 16(19.0)
32(76.1) 10(23.9)
36(85.7) 6(14.3)
0.266
103(41.5) 115(46.4) 30(12.1)
58(45.7) 50(39.3) 19(15.0)
45(37.2) 65(53.7) 11(9.1)
0.061
35(41.6) 38(45.3) 11(13.1)
19(45.2) 20(47.6) 3(7.1)
16(38.2) 18(42.8) 8(19.0)
0.268
152(61.3) 89(35.9) 7(2.8)
73(57.4) 50(39.3) 4(3.3)
79(65.3) 39(32.3) 3(2.4)
48(57.2) 34(40.5) 2(2.3)
23(54.7) 18(42.8) 1(2.5)
25(59.3) 16(38.2) 1(2.5)
0.904
194(78.3) 54(21.7)
103(81.1) 24(18.9)
91(74.3) 30(24.7)
61(72.6) 23(27.4)
32(76.2) 10(23.8)
29(69.0) 13(31.0)
132(53.0) 56(23.0) 60(24.0)
71(55.9) 34(26.8) 22(17.3)
67(55.4) 25(20.7) 29(23.9)
0.316
52(62.0) 15(17.8) 17(20.2)
25(59.5) 7(16.6) 10(23.9)
27(64.4) 8(19.0) 7(16.6)
0.714
1 (0-7)
1 (0 - 5)
2 (0 - 7)
<0.001
1.5 (0-7)
1 (0-5)
2 (0-7)
0.431
0.473
0.261
0.463
PSM: propensity score matching; pts: patients; FBT400: fludarabine (50 mg/m2/day for 4 days), busulfan (3.2 mg/kg/day for 4 days) and total body irradiation (TBI) 400 cGy; FBT200: fludarabine (30mg/m2/day for 4 days), busulfan (3.2 mg/kg/day for 2 days), and total body irradiation (TBI) 200 cGy; AML: Acute myeloid Leukemia; MDS: myelodysplastic syndrome; MRD: matched related donor; MUD: matched unrelated donor; MMUD: mismatched unrelated donor; GVHD: graft-versus-host disease; CSA:
cyclosporine; MMF: mycophenolate
mofetil; TCD: in vivo T-cell depletion; CR: complete remission; HCT-CI: hematopoietic cell transplantationcomorbidity index.
17 Page 17 of 22
Table 2. Comparison of outcomes between FBT 400 and FBT200.
Overall population Acute GVHD grade 1-4 grade 2-4 grade 3/4 Chronic GVHD at 1 y (any grade) Outcomes at 2 y OS NRM Relapse
PSM group
Overall % ± SE
FBT400 (n=127)
FBT200 (n=121)
p value
Overall % ± SE
FBT400 (n=42)
FBT200 (n=42)
p value
59.8±5.9 43.6±4.3 19.4±1.9
63.8±6.3 46.5±4.6 21.3±2.1
55.6±5.5 40.6±4.0 17.4±1.7
0.171 0.086 0.479
61.9±6.0 41.7±4.1 16.7±1.7
71.4±6.8 47.6±4.6 21.3±2.1
52.4±5.1 35.7±3.5 17.4±1.7
0.143 0.091 0.479
53.9±5.3
51.4±5.0
56.5±5.5
0.605
52.1±5.1
39.3±3.9
64.8±6.2
0.138
48.0±3.6 34.6±3.6 24.8±3.5
51.7±5.2 34.1±4.6 14.1±1.4
45.2±5.0 28.7±2.8 26.1±2.6
0.541 0.368 0.033
48.0±3.6 34.6±3.6 24.8±3.5
54.7±7.7 33.3±2.8 9.5±1.1
49.5±9.1 22.2±2.3 25.7±2.6
0.718 0.238 0.315
PSM: propensity score matching; GVHD: graft-versus-host disease; y: year; OS: overall survival; NRM: nonrelapse mortality.
18 Page 18 of 22
Table 3. UVA and MVA of 2-year OS in overall population and PSM group
Overall population UVA Parameter Age (years) Disease state at HSCT Donor type
Diagnosis Conditioning Cytogenetics HCT-CI Period effect
CR1 CR2 No CR MRD MUD MMUD AML MDS FBT400 FBT200 Fav/Interm Adverse ordinal continuous
MVA
UVA
PSM group MVA
HR [95% CI]
p value
HR [95% CI]
p value
HR [95% CI]
p value
HR [95% CI]
1.023 [1.005-1.042] 1.000 1.834 [1.096-2.816]
0.014 0.007
1.031 [1.011-1.051] 1.000 1.424 [1.0960-1.714]
0.001 0.018
1.012 [0.961-1.065] 1.000 1.013 [0.404-2.540]
0.640 0.978
1.012 [0.950-1.078] 1.000 1.300 [0.493-3.428]
0.710 0.595
1.000 2.500 [1.490-4.201]
0.002
-
-
1.000 1.508 [0.719-3.161]
0.002
-
1.000 1.628 [1.094-2.423] 1.000 1.117 [1.091-3.703] 1.000 1.628 [1.091-2.427] 1.124 [1.009-1.252] 1.020 [0.921-1.131]
0.015
1.000 1.464 [0.920-2.331] 1.000 1.606 [1.054-2.448] -
-
1.000 2.659 [1.328-5.319] 1.000 1.124 [0.594-2.132] 1.000 3.485 [1.796-6.761] 0.998 [0.829-1.202] 1.134 [0.931-1.381]
0.006
1.000 0.670[0.3427 -1.311] 1.000 2.574 [1.467-5.969] -
0.541 0.016 0.155 0.243
0.107 0.027 -
0.718 <0.001 0.980 0.211
p value
0.240 0.002 -
PSM: propensity score matching; UVA: univariate analysis; MVA: multivariate analysis; HR: hazard ratio; HSCT: hematopoietic stem cell transplant (allogeneic); CR: complete remission; MRD: matched related donor; MUD: matched unrelated donor; MMUD: mismatched unrelated donor; AML: Acute myeloid Leukemia; MDS: myelodysplastic syndrome; FBT400: fludarabine (50 mg/m2/day for 4 days), busulfan (3.2 mg/kg/day for 4 days) and total body irradiation (TBI) 400 cGy; FBT200: fludarabine (30mg/m2/day for 4 days), busulfan (3.2 mg/kg/day for 2 days), and total body irradiation (TBI) 200 cGy; Fav/interm: favorable/intermediate; HCT-CI: hematopoietic cell transplantation-comorbidity index.
19 Page 19 of 22
Table 4. UVA and MVA of NRM in Overall population and PSM group
Overall population UVA Parameter Age (years) Disease state at HCT Donor type
Diagnosis Conditioning Cytogenetics HCT-CI Period effect
HR [95% CI] CR1 CR2 No CR MRD MUD MMUD AML MDS FBT400 FBT200 Fav/Interm Adverse ordinal continuous
p value
1.012 [0.990-1.033] 1.000 1.465 [1.125-1.906]
0.280 0.004
1.000 1.79[1.320-2.427]
0.001
1.000 2.183 [1.353-3.521] 1.000 0.855[0.544-1.344] 1.000 1.385[0.809-2.372] 1.117[0.971-1.284] 1.064[0.931-1.215]
0.001 0.500 0.230 0.120 0.360
MVA HR [95% CI] 1.000 1.759 [1.298-2.384] 1.000 2.044 [1.256-3.322] 1.000 0.806 [0.495-1.314] -
UVA p value
HR [95% CI]
PSM group MVA p value
HR [95% CI]
p value
-
0.992 [0.926-1.064] 1.000 1.919 [1.186-3.104]
0.840 0.008
1.000 1.317 [0.547-3.166]
0.540
0.002
1.000 2.195 [1.302-3.701]
0.003
1.000 2.144 [1.085-4.236]
0.028
0.004
1.000 0.306 [0.133-0.701] 1.000 0.609 [0.267-1.390] 1.000 2.570 [1.132-5.839] 0.749 [0.544-1.031] 1.266 [0.989-1.621]
0.005
1.000 0.650 [0.125-3.378] 1.000 0.516 [0.206-1.293] -
0.610
0.390 -
0.240 0.024 0.077 0.061
0.160 -
PSM: propensity score matching; UVA: univariate analysis; MVA: multivariate analysis; HR: hazard ratio; HSCT: hematopoietic stem cell transplant (allogeneic); CR: complete remission; MRD: matched related donor; MUD: matched unrelated donor; MMUD: mismatched unrelated donor; AML: Acute myeloid Leukemia; MDS: myelodysplastic syndrome; FBT400: fludarabine (50 mg/m2/day for 4 days), busulfan (3.2 mg/kg/day for 4 days) and total body irradiation (TBI) 400 cGy; FBT200: fludarabine (30mg/m2/day for 4 days), busulfan (3.2 mg/kg/day for 2 days), and total body irradiation (TBI) 200 cGy; Fav/interm: favorable/intermediate; HCT-CI: hematopoietic cell transplantation-comorbidity index.
20 Page 20 of 22
Table 5. UVA and MVA of CIR in Overall population and PSM group.
Overall population UVA Parameter Age (years) Disease state at HCT
HR [95% CI]
p value
MVA HR [95% CI]
UVA p value
HR [95% CI]
PSM group MVA p value
HR [95% CI]
CR1 CR2
1.039 [1.006-1.073] 1.000 1.035 [0.755-1.416]
0.021 0.830
1. 045 [1.011-1.081] 1.000 1.833 [1.098-3.061]
0.009 0.020
1.039 [0.969-1.113] 1.000 1.116 [0.625-1.992]
0.280 0.710
1.005 [0.918-1.100] 1.000 1.190 [0.429-3.303]
Donor type
No CR MRD MUD
1.000 0.834 [0.531-1.311]
0.43
-
-
1.000 1.089 [0.478-2.518]
0.850
-
Diagnosis
MMUD AML MDS
1.000 1.437 [0.698-2.954]
0.353
0.006
1.000 1.802 [0.429-7.571]
0.42
FBT400 FBT200 Fav/Interm Adverse ordinal continuous
1.000 1.932 [1.087-3.433] 1.000 1.675 [0.917-3.057] 1.437 [0.6986-2.954] 0.9232 [0.7962-1.07]
0.025
1.000 4.148 [1.49411.520] 1.000 1.341 [0.662-2.714] 1.000 2.053 [1.083-3.890] -
0.410
1.000 1.683 [0.598-4.729] 1.000 2.831 [1.047-7.657] 1.387 [1.131-1.700] 0.793 [0.588-1.071]
0.32
Conditioning Cytogenetics HCT-CI Period effect
0.093 0.320 0.290
0.027 -
0.04 0.001 0.130
1.000 4.236 [0.627-28.610] 1.000 1.428 [0.448-4.552] 1.000 3.813 [1.162-12.510] -
p value 0.910 0.740
0.140
0.550 0.027 -
PSM: propensity score matching; UVA: univariate analysis; MVA: multivariate analysis; HR: hazard ratio; HSCT: hematopoietic stem cell transplant (allogeneic); CR: complete remission; MRD: matched related donor; MUD: matched unrelated donor; MMUD: mismatched unrelated donor; AML: Acute myeloid Leukemia; MDS: myelodysplastic syndrome; FBT400: fludarabine (50 mg/m2/day for 4 days), busulfan (3.2 mg/kg/day for 4 days) and total body irradiation (TBI) 400 cGy; FBT200: fludarabine (30mg/m2/day for 4 days), busulfan (3.2 mg/kg/day for 2 days), and total body irradiation (TBI) 200 cGy; Fav/interm: favorable/intermediate; HCT-CI: hematopoietic cell transplantation-comorbidity index.
21 Page 21 of 22
Figure 1.
22 Page 22 of 22