Autologous Is Superior to Allogeneic Hematopoietic Cell Transplantation for Acute Promyelocytic Leukemia in Second Complete Remission

(1)

Autologous Is Superior to Allogeneic

Hematopoietic Cell Transplantation for Acute

Promyelocytic Leukemia in Second Complete

Remission

Jennifer L. Holter Chakrabarty

1,*

_{, Morel Rubinger}

2

_,

Jennifer Le-Rademacher

3

, Hai-Lin Wang

4

, Andrew Grigg

5

,

George B. Selby

1

, Jeffrey Szer

6

, Jacob M. Rowe

7

,

Daniel J. Weisdorf

8

, Martin S. Tallman

9

1_{Division of Hematology, Oncology and Transplantation, Department of Medicine, University of}

Oklahoma, Oklahoma City, Oklahoma

2_{Department of Hematology, CancerCare Manitoba, Winnipeg, Canada}

3_{Divison of Biostatistics, Center for International Blood and Marrow Transplant Research,}

Medical College of Wisconsin, Milwaukee, Wisconsin

4_{Center for International Blood and Marrow Transplant Research, Medical College of Wisconsin,}

Milwaukee, Wisconsin

5_{Department Clinical Haematology, Austin Hospital, Melbourne, Australia}

6_{Department Clinical Haematology and Bone Marrow Transplantation, Royal Melbourne Hospital City}

Campus, Victoria, Australia

7_{Department of Hematology and Oncology, Rambam Medical Center, Haifa, Israel}

8_{Center for International Blood and Marrow Transplant Research, University of Minnesota Medical}

Center, Minneapolis, Minnesota

9_{Department of Medicine, Leukemia Service, Memorial Sloan Kettering Cancer Center, New York, New}

York Article history: Received 19 December 2013 Accepted 24 March 2014 Key Words: APL Autologous transplantation Allogeneic transplantation

a b s t r a c t

To identify favored choice of transplantation in patients with acute promyelocytic leukemia (APL) in second complete remission, we studied 294 patients with APL in second complete remission (CR2) receiving allo-geneic (n¼ 232) or autologous (n ¼ 62) hematopoietic cell transplantation (HCT) reported to the Center for International Blood and Marrow Transplantation Research (CIBMTR) from 1995 to 2006, including 155 with pre-HCT PML/RARf status (49% of allogeneic and 66% of autologous). Patient characteristics and trans-plantation characteristics, including treatment-related mortality, overall survival (OS), and disease-free sur-vival, were collected and analyzed for both univariate and multivariate outcomes. With median follow-up of 115 (allogeneic) and 72 months (autologous), 5-year disease-free survival (DFS) favored autologous with 63% (49% to 75%), compared with allogeneic at 50% (44% to 57%) (P¼ .10). OS was 75% (63% to 85%) versus 54% (48% to 61%) (P¼ .002), for autologous and allogeneic transplantation, respectively. Multivariate analysis showed significantly worse DFS after allogeneic HCT (hazard ratio [HR], 1.88; 95% confidence interval [CI], 1.16 to 3.06; P¼ .011) and age > 40 years (HR, 2.30; 95% CI, 1.44 to 3.67; P ¼ .0005). OS was significantly worse after allogeneic HCT (HR, 2.66; 95% CI, 1.52 to 4.65; P¼ .0006); age > 40 (HR, 3.29; 95% CI, 1.95 to 5.54; P< .001), and first complete remission < 12 months (HR, 1.56; 95% CI, 1.07 to 2.26; P ¼ .021). Positive pre-HCT PML-RARf status in 17 of 114 allogeneic and 6 of 41 receiving autologous transplantation did not influence relapse, treatment failure, or survival in either group. The survival advantage for autografting was attributable to increased treatment-related mortality (TRM) in the allogeneic group of 30% compared to 2% in the autologous group, in addition to the added mortality associated with GVHD. We conclude that autologous HCT yields superior OS for APL in CR2. Long-term DFS in autologous recipients, even with minimal residual diseaseepositive grafts, remains an important subject for further study.

Ó 2014 American Society for Blood and Marrow Transplantation.

INTRODUCTION

Acute promyelocytic leukemia (APL) accounts for 10% to 15% of de novo acute myeloid leukemia in younger adults[1]. Therapeutic strategies of all-trans retinoic acid, anthracycline-based induction, and recently, arsenic trioxide (ATO) yield remission rates of approximately 90%, most of which are du-rable with 2-year disease-free survival (DFS) of 79% and overall survival (OS) of 92%[2,3]. For the 10% to 20% of patients who relapse, recent evidence suggests that ATO reinduction can

lead to second complete remission (CR2) in up to 85% of relapsed, previously arsenic-naïve patients [4,5]. Although hematopoietic cell transplantation (HCT) is the generally accepted therapy for APL in CR2, the choice of allogeneic versus autologous HCT remains controversial. It is uncer-tain whether the increased treatment-related mortality (TRM) generally associated with allografts is compensated for by a lower relapse rate due to a graft-versus-APL effect. It has also been suggested that the outcome for autologous HCT is best if molecularly negative cells are collected[6,7], but the relative impact of persistent PML-RAR_{f transcripts} before allograft or autologous HCT on outcome is uncer-tain. In this retrospective study, we analyzed allogeneic

Financial disclosure: See Acknowledgments on page 1024.

* Correspondence and reprint requests: Jennifer L. Holter Chakrabarty, MD, 800 NE 10th Street, Suite 6023, Oklahoma City, OK 73104.

E-mail address:jholter2@ouhsc.edu(J.L. Holter Chakrabarty).

1083-8791/$e see front matter Ó 2014 American Society for Blood and Marrow Transplantation. http://dx.doi.org/10.1016/j.bbmt.2014.03.025

American Society for Blood

ASBMT

and Marrow Transplantation

(2)

and autologous HCT in CR2 to compare their toxicities and survival outcomes and to evaluate the impact of residual molecularly detectable pretransplantation marrow disease on outcome.

METHODS

This multi-institutional, international, retrospective study used data on autologous and allogeneic HCT for APL in CR2 reported to the Center for International Blood and Marrow Transplant Research (CIBMTR) from 79 centers in 18 countries during the years 1995 to 2006. Variables analyzed for their influence on outcome included age, time from first complete remission (CR1) to relapse (< or >12 months), and, for allografts, type of donor. In addition to the information available from standard CIBMTR report forms, supplemental data were retrieved from the HCT centers to determine mo-lecular and cytogenetic results before HCT, time from second complete remission (CR2) to transplantation, and therapy used to attain CR2. Evidence of positivity of disease status just before transplantation defined minimal residual disease (MRD). Supplemental forms collected both quantitative and qualitative molecular and cytogenetic data for establishment of MRD. Sup-plemental data were retrieved in 114 of 232 (49%) of allogeneic and 41 of 62 (66%) of autologous cases.

Clinical and demographic characteristics of the 2 groups were examined using chi-square statistics. Additional analyses compared the characteristics of patients with or without the requested supplemental data to insure generalizability of data from the smaller cohort with this added detail available. Defined endpoints included DFS as the time from HCT to death or relapse. Relapse was defined as time to hematologic recurrence and TRM was defined as death within the first 21 days or death before relapse. Overall survival (OS) was calculated as the time from HCT to death or last contact alive. Statistical analysis was performed using proportional hazards regression models. The proportional hazards assumptions for all variables were examined and all variables met the assumptions. A stepwise selection procedure was used to select variables significantly associated with the outcomes. Because the type of transplantation was the main effect of in-terest, it was included in all steps of the model building process. Risk factors with a significance level of .05 were included in the final models. The interaction between the main effect of transplantation type and all signifi-cant covariates was examined and no interaction was found to be statisti-cally significant. All analyses were conducted using SAS V9.3 (SAS Institute, Cary, NC).

RESULTS

Patient characteristics and transplantation-related factors

are shown in Table 1. The numbers of allogeneic HCT

decreased over time, whereas the relative number of autol-ogous HCT increased in more recent years.

Patient demographics, including Karnofsky performance status and sex, were not statistically different between allogeneic and autologous recipients. However, allogeneic recipients were significantly younger. Disease-related factors including histologic subtype, white blood cell (WBC) count at diagnosis, cytogenetic and/or molecular positivity at diag-nosis, time from CR1 to relapse, and time from CR2 to transplantation were not significantly different. The use of ATO occurred during the period of study, representing a major change in the treatment of APL. For this reason, we evaluated the impact of ATO therapy before transplantation. In univariate and multivariate analysis, we observed no impact of ATO-containing versus non-ATO pre-HCT therapy on the risk of relapse after HCT (hazard ratio [HR], .598; 95% confidence interval [CI], .230 to 1.456; P ¼ .25).

The vast majority of autologous transplantations were myeloablative, using standard regimens, most often total body irradiation (TBI) (Table 1). Univariate outcomes are shown inTable 2, with favorable OS, DFS, and TRM observed

in autologous HCT. Only a modest and not signiﬁcant

reduction in relapse followed allogeneic HCT. Five-year OS for the autologous cohort was 75% (95% CI, 63% to 85%) and for the allogeneic cohort, 54% (95% CI, 48% to 61%) (Figure 1). The risk of overall mortality (inverse of OS) was higher after allogeneic transplantation compared with autologous (HR,

2.66; 95% CI, 1.52 to 4.65; P ¼ .0006). DFS at 5 years also favored autologous HCT: 63% (49% to 75%) compared with 50% (44% to 57%) in allogeneic cohort (Figure 2). Equivalently, the risk of treatment failure (inverse of DFS) was higher after allogeneic transplantation (HR, 1.88; 95% CI, 1.16 to 3.06; P¼ .011). Multivariate analysis showed that age > 40 years also adversely inﬂuenced overall mortality (HR, 3.29; 95% CI, 1.95 to 5.54; P<.001) and treatment failure, with a HR of 2.30 (95% CI, 1.44 to 3.67); P¼ .0005. A shorter duration of CR1 (< 12 months) was associated with signiﬁcantly higher mortality (HR, 1.56; 95% CI, 1.07 to 2.26; P¼ .02).

The molecular status of PML-RARf transcripts or

cy-togenetic translocations detectable in the pre-HCT bone

marrow were studied for their inﬂuence on relapse and

other HCT outcomes. The demographics of the cohorts

with these additional data are shown in supplemental

Table S1. Surprisingly, in both autologous (n ¼ 6 of 41) and allogeneic HCT (n¼ 17 of 114), molecular or cytoge-netically positive grafts were not associated with increased

Table 1

Patient and Transplantation Characteristics

Patient Characteristics Allogeneic Autologous P Value Patients, n 232 62 Age at transplantation, n (%) < .001 <20 years 70 (30) 9 (15) 20-40 years 93 (40) 19 (31) >40 years 69 (30) 34 (55) Year of HCT < .001 1995-2000 145 (63) 25 (40) 2001-2006 87 (38) 37 (60) Sex .18 Male 116 (50) 37 (60) Female 116 (50) 25 (40) KPS .91 > 90% 178 (77) 46 (74) < 90% 48 (21) 14 (23) Data missing 6 (3) 2 (3)

Time from CR1 to relapse, median, (range), mo 14 (< 1-87) 17 (< 1-85) .08 <12 months 83 (36) 18 (29) >12 Months 130 (56) 33 (53) Data missing 19 (8) 11 (18) Time from CR2 to HCT, median (range), mo 2 (< 1-45) 2 (< 1-10) .19 <6 months 202 (87) 59 (95) >6 months 28 (12) 3 (5) Data missing 2 (< 1) – Conditioning regimen Myeloablative 213 (92) 55 (89) RIC 15 (7) 5 (8) Data missing 4 (2) 2 (4) TBI containing 116 (50) 47 (76) .001 Non-TBI 114 (49) 15 (24)

Donor type N/A

HLA-identical Sibling 124 (53) URD well/partially matched 63 (27) Other related 45 (19)

Stem cell source < .001

Marrow 154 (66) 8 (12)

Peripheral blood 78 (34) 54 (88)

GVHD prophylaxis N/A

CSA/Tac MMF 203 (88) T cell depletionþ post-HCT

immunosuppression

24 (10)

Other 5 (2)

KPS indicates Karnofsky performance score (before conditioning); RIC, reduced-intensity conditioning; TBI, total body irradiation; URD, unrelated donor; GVHD, graft-versus-host disease; CSA, cyclosporine; Tac, tacrolimus; MMF, mycophenolate mofetil; CR1,ﬁrst complete remission; CR2, second complete remission; HCT, hematopoietic cell transplantation.

(3)

risks. The 3-year risks of relapse for autologous positive pre-HCT was 17% (95% CI, 0 to 55%) versus negative grafts, 24% (95% CI, 10% to 41%); P¼ .69; for allogeneic positive, 27% (95% CI, 8% to 51%) versus negative, 16% (95% CI, 9% to

25%); P¼ .38. DFS with autologous positive pre-HCT was

83% (95% CI, 46% to 100%) versus negative grafts, 69% (95% CI, 51% to 84%); P¼ .41; allogeneic positive pre-HCT, 60% (95% CI, 35% to 82%) versus negative, 62% (95% CI, 51% to 73%); P¼ .87. Overall survival with autologous positive pre-HCT was 83% (95% CI, 46% to 100%) versus negative grafts, 77% (95% CI, 60% to 90%); P¼ .70; allogeneic positive, 65% (95% CI, 41% to 85%) versus negative, 66% (95% CI, 55% to 76%); P¼ .90.

Multivariate analysis confirmed that positive molecular or cytogenetic status was not significantly associated with overall mortality (positive, HR, .73; 95% CI, .318 to 1.651; P¼ .44). In the allogeneic group with positive molecular or cytogenetically detectable disease before HCT, the relapse rate at 5 years was 27% (95% CI, 8% to 51%), DFS was 60% (95% CI, 35% to 82%), and OS was 65% (95% CI, 41% to 85%). All were similar to these corresponding outcomes in those who had negative pre-HCTfindings. Notably, in the autologous group,

only 6 patients had positive cytogenetic or molecular testing before HCT, yet of these 6, 1 relapsed at 6 months and died 11 months after HCT. Five remain alive and disease free at 47, 60, 61, 72, and 129 months after HCT. We acknowledge the small numbers of patients with positive grafts makes any ﬁrm conclusions difﬁcult.

In the allogeneic group, the cumulative incidence of acute GVHD grade II to IV at 100 days was 33% (95% CI, 27% to 40%), whereas chronic GVHD at 5 years was 45% (95% CI, 38% to 52%), consistent with established rates in allogeneic HCT. Figure 3shows the cumulative incidence of TRM in the 2 cohorts. Causes of death for all patients (123 of 294) were

primarily leukemia recurrence (n¼ 33), new malignancy

(n¼ 4), idiopathic pneumonia syndrome (n ¼ 17), infection (n_{¼ 17), and organ failure (n ¼ 21). The causes of death in} allogeneic transplantation group (n¼ 107 of 232) were less often APL recurrence (n¼ 27), but more often due to toxic-ities, including organ toxicity/failure (n¼ 18), GVHD (n ¼ 16), infection (n¼ 15), idiopathic pneumonia syndrome (n ¼ 15), and others (2 graft failure, 2 hemorrhage, 1 thromboembolic, and 6 other not speciﬁed).

Figure 1. Overall survival in patients undergoing allogeneic and autologous transplantation with APL in CR2.

Figure 2. Disease-free survival in patients undergoing allogeneic and autolo-gous transplantation for APL in CR2.

Figure 3. Treatment-related mortality in patients undergoing allogeneic and autologous transplantation for APL in CR2.

Table 2

Univariate Probability of Outcome

Outcomes Allogeneic (n¼ 232) Autologous (n¼ 62) P Value n Prob (95% CI), % n Prob (95% CI), % Relapse 230 61 1 year 10 (6-14) 21 (11-32) .14 3 year 16 (12-22) 28 (17-40) .34 5 year 18 (14-24) 30 (19-42) .40 Treatment-related mortality 230 61 1 year 28 (22-34) 2 (0-8) < .001 3 year 30 (24-36) 5 (1-14) < .001 5 year 31 (25-37) 7 (2-17) < .001 Disease-free survival 230 61 1 year 62 (56-68) 78 (66-87) .02 3 year 54 (47-60) 67 (54-78) .07 5 year 50 (44-57) 63 (49-75) .10 Overall survival 232 62 1 year 66 (60-72) 93 (85-98) < .001 3 year 58 (52-65) 79 (67-88) .002 5 year 54 (48-61) 75 (63-85) .002

(4)

DISCUSSION

This large, multinational series demonstrated that the preferred therapy for APL in CR2 is autologous over alloge-neic HCT. The signiﬁcantly greater TRM of the allogeneic HCT is not overcome by the modest but not signiﬁcantly better protection against relapse accompanying the allogeneic graft process, even when adjusted for the molecular or cytogenetic status of the patient before HCT in CR2[7]. An earlier analysis by the European Cooperative Group for Blood and Marrow Transplantation reviewed 625 patients with APL who underwent transplantation by either autologous or allogeneic HSCT in either CR1 or CR2 from

1993 to 2003 [8]. Patients in CR2 (n ¼ 195) had lower

relapse rates after allogeneic HCT (37% 4% autologous

versus 17% 3%) and 5-year leukemia-free survival of 51% 4% autologous versus 59% 4% after allografts. Autolo-gous transplantations for APL in this study had a higher

TRM of 16% 3%. The higher than expected TRM in the

autologous arm led to a small difference of TRM between

allogeneic transplants (24% 4%). Allogeneic HCT was

recommended in CR2 when a sibling donor was available [8].

Few data on molecular-positive patients who underwent transplantation during CR2 are available. Meloni et al. analyzed 15 patients with relapsed APL undergoing autolo-gous HCT, and PCR-negative pre-HCT marrow led to 45% (6 of 15) of patients remaining in remission[9]. However, 7 of 7 patients with PCR-positive marrow relapsed within 9 months of autologous HCT. They recommended allogeneic HCT if the pre-HCT marrow was cytogenetically or molecu-larly positive. Sanz et al. described a single patient in CR2 with molecular positivity on preconditioning marrow who received an autologous HCT and survived leukemia free 22 months after HCT[10].

After_{ﬁrst relapse, several highly effective strategies for} APL include ATO for reinduction and transplantation dur-ing CR2 [11]. Our data highlight that, despite continued improvements in supportive care and improvements in management of GVHD, the substantial excess of TRM limits survival for allogeneic HCT patients with APL in CR2. Somewhat surprisingly, pretransplantation molecular and cytogenetic positivity identifying detectable residual

dis-ease at HCT had no signiﬁcant inﬂuence on treatment

failure. In contrast to theﬁndings of Meloni et al., and in an admittedly small cohort, residual detectable disease rarely led to failure in the autologous group with 5 of 6 patients remaining disease free. This observation suggests that effective eradication of residual disease in vivo, preferen-tial mobilization into the autograft of short-term repopu-lating cells, but too few leukemia stem cells to induce

relapse [12], a modest purging effect of

cryopreser-vation on unstable leukemic clones or other mechanisms may be protective and offers opportunities for further investigation.

The remarkable success obtained with the newest strategy, all-trans retinoic acid plus ATO for initial therapy, will likely lead to less relapse and decrease in the need for

any kind of transplantation [13]. However, these data

further clarify that APL exhibits unique characteristics, making it amenable to autologous HCT, compared with other types of leukemia, where allogeneic HCT may be preferred. These data provide evidence that autologous HCT for APL is the superior treatment strategy, leading to long-term remission and survival for a large proportion of patients in CR2.

ACKNOWLEDGMENTS

CIBMTR funding support: The CIBMTR is supported by Public Health Service Grant/Cooperative Agreement U24-CA076518 from the National Cancer Institute, the National Heart, Lung, and Blood Institute and the National Institute of Allergy and Infectious Diseases; a Grant/Cooperative Agreement 5U10HL069294 from NHLBI and NCI; a contract HHSH250201200016C with Health Resources and Services Administration; two Grants 12-1-0142 and

N00014-13-1-0039 from the Ofﬁce of Naval Research; and grants

from *Actinium Pharmaceuticals; Allos Therapeutics, Inc.;

*_{Amgen, Inc.; Anonymous donation to the Medical College of}

Wisconsin; Ariad; Be the Match Foundation;*Blue Cross and Blue Shield Association; *Celgene Corporation; Chimerix, Inc.; Fred Hutchinson Cancer Research Center;

Fresenius-Biotech North America, Inc.; *Gamida Cell Teva Joint

Venture Ltd.; Genentech, Inc.;*Gentium SpA; Genzyme Cor-poration; GlaxoSmithKline; Health Research, Inc. Roswell Park Cancer Institute; HistoGenetics, Inc.; Incyte Corpora-tion; Jeff Gordon Children’s Foundation; Kiadis Pharma; The Leukemia & Lymphoma Society; Medac GmbH; The Medical College of Wisconsin; Merck & Co, Inc.; Millennium: The Takeda Oncology Co.;*Milliman USA, Inc.;*Miltenyi Biotec, Inc.; National Marrow Donor Program; Onyx Pharmaceuti-cals; Optum Healthcare Solutions, Inc.; Osiris Therapeutics, Inc.; Otsuka America Pharmaceutical, Inc.; Perkin Elmer, Inc.;

*_{Remedy Informatics;}*_Sano_{ﬁ US; Seattle Genetics;}

Sigma-Tau Pharmaceuticals; Soligenix, Inc.; St. Baldrick’s Founda-tion; StemCyte, A Global Cord Blood Therapeutics Co.; Stemsoft Software, Inc.; Swedish Orphan Biovitrum;*Tarix Pharmaceuticals; *TerumoBCT; *Teva Neuroscience, Inc.;

*THERAKOS, Inc.; University of Minnesota; University of Utah; and*Wellpoint, Inc. The views expressed in this article do not reﬂect the ofﬁcial policy or position of the National Institute of Health, the Department of the Navy, the Department of Defense, Health Resources and Services Administration, or any other agency of the US Government. *Corporate Members.

Financial disclosure: The authors have nothing to disclose. Conﬂict of interest statement: There are no conﬂicts of in-terest to report.

SUPPLEMENTARY DATA

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.bbmt.2014.03.025.

REFERENCES

1. Nabhan C, Mehta J, Tallman MS. The role of bone marrow trans-plantation in acute promyelocytic leukemia. Bone Marrow Transplant. 2001;28:219-226.

2. Tallman MS. Arsenic trioxide: its role in acute promyelocytic leukemia and potential in other hematologic malignancies. Blood Rev. 2001;15: 133-142.

3. Mandelli F, Diverio D, Awisati G, et al. Molecular remission in PML/RAR alpha-positive acute promyelocytic leukemia by combined all-trans retinoic acid and idarubicin (AIDA) therapy. Gruppo Italiano-Malattie Ematologiche Maligne dell’Adulto and Associazione Italiana di Emato-logia ed OncoEmato-logia Pediatrica Cooperative Groups. Blood. 1997;90: 1014-1021.

4. Soignet SL, Frankel SR, Douer D, et al. United States multicenter study of arsenic trioxide in relapsed acute promyelocytic leukemia. J Clin Oncol. 2001;19:3852-3860.

5. Soignet SL, Maslak P, Wang ZG, et al. Complete remission after treat-ment of acute promyelocytic leukemia with arsenic trioxide. N Engl J Med. 1998;339:1341-1348.

6. Jurcic JG, Nimer SD, Scheinberg DA, et al. Prognostic signiﬁcance of minimal residual disease detection and PML/RAR-alpha isoform type: long-term follow-up in acute promyelocytic leukemia. Blood. 2001;98: 2651-2656.

(5)

7. Grimwade D, Tallman MS. Should minimal residual disease monitoring be the standard of care for all patients with acute promyelocytic leu-kemia? Leuk Res. 2011;35:3-7.

8. Sanz MA, Labopin M, Gorin NC, et al. Hematopoietic stem cell trans-plantation for adults with acute promyelocytic leukemia in the ATRA era: a survey of the European Cooperative Group for Blood and Marrow Transplantation. Bone Marrow Transplant. 2007;39:461-469. 9. Meloni G, Diverio D, Vignetti M, et al. Autologous bone marrow

transplantation for acute promyelocytic leukemia in second remission: prognostic relevance of pretransplant minimal residual disease assessment by reverse-transcription polymerase chain reaction of the PML/RAR alpha fusion gene. Blood. 1997;90:1321-1325.

10.Sanz MA, de la Rubia J, Bonanad S, et al. Prolonged molecular remission after PML/RAR alpha-positive autologous peripheral blood stem cell transplantation in acute promyelocytic leukemia: is relevant pretrans-plant minimal residual disease in the graft? Leukemia. 1998;12:992-995. 11.Yanada M, Tsuzuki M, Fujita H, et al. Phase 2 study of arsenic trioxide followed by autologous hematopoietic cell transplantation for relapsed acute promyelocytic leukemia. Blood. 2013;121:3095-3102. 12.Wang JC, Dick JE. Cancer stem cells: lessons from leukemia. Trends Cell

Biol. 2005;15:494-501.

13.Estey E, Garcia-Manero G, Ferrajoli A, et al. Use of all-trans retinoic acid plus arsenic trioxide as an alternative to chemotherapy in untreated acute promyelocytic leukemia. Blood. 2006;107:3469-3473.