Secondary acute myeloid leukemia: A risk factor for allogeneic stem cell transplantation in first complete remission

Secondary acute myeloid leukemia (sAML) evolves from underlying hematological conditions, such as myelodysplastic syndrome (MDS), myeloproliferative neoplasms (MPN) and bone marrow failure disorders. sAML can also arise following subjection to radio- or chemotherapy, resulting in a unique subtype known as therapy related AML (tAML).¹

Relative to de novo AML, the prognosis of sAML is poor, and patients respond inadequately to conventional chemotherapy and autologous stem cell transplants (auto-SCT). The front-line therapy for sAML is allogenic SCT (allo-SCT), however many patients lack human leukocyte antigen identical siblings.²

It remains unanswered whether sAML is, itself, a contributor to poor outcomes following allo-SCT. The Acute Leukemia Working Party (ALWP) of the European Society of Blood and Marrow Transplantation (EBMT) aimed to address this question; conducting a retrospective comparison of patient outcomes following allo-SCT for the treatment of sAML vs de novo AML. Ann-Kristin Schmaelter and coinvestigators published the findings in Blood Cancer Journal and the AML Hub hereby presents a summary.²

Study design

• Data from over 600 transplant centers were obtained from the EBMT registry for retrospective analysis
• Adult patients receiving allo-SCT for the treatment of de novo (n = 11,439) or sAML (n = 1325) between January 2000 and December 2016 were eligible
• Allo-SCT origin
  • matched related
  • 9/10 or 10/10 antigen matched unrelated
  • T-cell replete haploidentical donor
• Primary outcomes: overall survival (OS), leukemia free survival (LFS), graft-versus-host disease (GvHD)/relapse free survival (GRFS), cumulative incidence of relapse (RI), non-relapse mortality (NRM), acute and chronic GvHD, and cause of death

Results

• Median follow-ups after allo-SCT were 36.1 months and 33.1 months for patients with de novo AML and sAML respectively

Patient characteristics

• Of the sAML cases, 62.5% arose from previous myeloid malignancy or bone marrow failure syndrome and 37.7% were tAML

Table 1. Baseline characteristics for patients receiving allo-SCT for the treatment of de novo AML and sAML

allo-SCT, allogenic stem cell transplantation; AML, acute myeloid leukemia; CR1/2, first/second complete remission; MAC, myeloablative conditioning; MSD, matched sibling donor; PIF, primary induction failure; RIC, reduced-intensity conditioning; sAML, secondary AML; SCT stem cell transplantation; UD, unrelated donor *Allele compatibility for the human leukocyte antigen (HLA) † Nine out of ten donor allele match to HLA
Characteristic, p < 0.001	De novo AML (n = 11,439)	sAML (n = 1325)
Median age at SCT, years (range)	49.3 (18–76.8)	57.7 (18.3–76)
Status at SCT, % CR1 CR2 PIF Relapse	67.23 18.64 5.31 8.82	68.6 7.02 15.02 9.36
Donor MSD UD 10/10* UD 9/10† Haploidentical donor	55.99 28.68 9.05 6.28	43.62 39.09 10.49 6.79
Cytogenetics Favorable Intermediate Adverse	13.04 68.27 18.69	4.38 63.09 32.53
Graft source Bone marrow Peripheral blood	17.53 82.47	12.23 87.77
Conditioning regimen MAC RIC	55.13 43.32	44.87 56.68

Patient outcomes

• The majority of patients reached engraftment following allo-SCT irrespective of diagnosis (Table 2)
• The most common cause of death was recurrence of original disease

Table 2. Patient outcomes to allo-SCT by diagnosis

allo-SCT, allogenic stem cell transplantation; AML, acute myeloid leukemia; aGvHD, acute graft-versus-host disease; cGvHD, chronic GvHD; GRFS, GvHD/relapse-free survival; LFS, leukemia free survival; OS, overall survival; RI, incidence of relapse; sAML, secondary AML; SCT stem cell transplantation
Outcome, %	De novo AML (n = 11,439)	sAML (n = 1325)
Engraftment	98.2	96.6
Patient outcome at 3 years OS LFS NRM RI GRFS	54.6 48.9 18 33 34.3	43.1 37.9 24.8 37.3 25.8
Grade 3–4 GvHD aGvHD cGvHD	23.5 40.6	24.4 35.4

Transplantation in complete remission 1 (CR1)

• As there was a significant interaction between diagnosis (de novo vs sAML) and disease status at the time of allo-SCT (Table 1), the impact of sAML was stratified by patient disease status at the time of allo-SCT
  • Most patients received SCT during CR1 ( Table 1)
  • A univariate analysis was conducted comparing the outcomes of patients transplanted in this period (Table 3)
  • sAML diagnosis had a negative impact on the 3-year patient outcomes (Table 3)

Table 3. The 3-year outcomes of patients receiving allo-SCT during CR1

AML, acute myeloid leukemia; CR1, first complete remission; GRFS, graft-versus-host disease (GvHD)/relapse-free survival; LFS, leukemia free survival; OS, overall survival; sAML, secondary AML
Outcome, %	De novo AML (n = 11,439)	sAML (n = 1325)	p
OS	60.8	46.7	< 0.00001
Relapse	28.5	35	0.000086
NRM	16.4	23.4	< 0.00001
LFS	55.1	41.6	< 0.00001
GRFS	38.6	28.4	< 0.00001

 

• A multivariate analysis of patients receiving allo-SCT during CR1 also concluded that sAML was associated with significantly reduced OS, LFS, GRFS and increased cumulative NRM and RI (Table 4)
  • This did not translate to patients receiving transplants during primary induction failure (PIF) or active relapse
• sAML was not a significant risk factor for GvHD progression
• A matched-pair analysis (n = 877 pairs) supplemented the findings; reinforcing that sAML significantly worsens patient outcomes to allo-SCT when compared to de novo AML

Table 4. Multivariate analysis of patients receiving allo-SCT during CR1

CI, confidence interval; GRFS, graft-versus-host disease (GvHD)/relapse-free survival; HR, hazard ratio; LFS, leukemia free survival; NRM, non-relapse mortality; OS, overall survival; RI, incidence
Outcome	HR	95% CI	p
OS	1.33	1.21–1.48	< 0.00001
LFS	1.32	1.19–1.45	< 0.00001
GRFS	1.20	1.10–1.31	< 0.0001
NRM	1.37	1.17–1.59	< 0.001
RI	1.27	1.12–1.44	< 0.0001

 Conclusion

• sAML, including tAML, was identified as an independent risk factor for OS, LFS, GRFS, RI, and NRM after allo-SCT during CR1
• Long-term remission of sAML is poor, irrespective of previous treatment and induction therapies
• The data from this study could be applied to improve transplant protocols, risk stratifications and prognostic valuations in patients receiving allo-HCT for sAML
• Limitations:
  • Where cytogenetic information was unavailable, patients were excluded from the analysis which introduced a selection bias
  • There was a significant lack of information on MRD status and pre- and post-transplant therapies in most patients
  • Patients receiving allo-SCT in the second remission were likely underrepresented in the matched-pair analysis