In the November 15, 2005, issue of Blood, Gale et al1  examined the role of stem cell transplantation (SCT) in patients with fetal liver tyrosine kinase 3/internal tandem duplication (FLT3/ITD)-positive acute myeloid leukemia (AML) and concluded that the presence of FLT3/ITD is not an indication for transplantation. We respectfully disagree with the authors, as the data presented in the article do not fully support their conclusion. Rather, the data suggest that allogeneic transplantation negates the poor prognostic significance of FLT3/ITD.

This article compared “autograft” versus “no SCT,” “autograft” versus “allograft,” as well as “donor” versus “no donor” for patients with and without FLT3/ITD. However, there is an important omission of a direct comparison of allograft recipients versus those who were treated with chemotherapy only. Based on the data Gale et al1  provided in Tables 4 and 5 (summarized here in Table 1), chemotherapy-only recipients have a worse outcome, with relative hazard ratios (RHRs) of 5.9 for relapse and 3 for death, compared with the “donor” group, suggesting that allograft may improve outcome in FLT3/ITD-positive AML.

Table 1.

Relative hazard ratios for data from Gale et al1



FLT3/ITD positive

FLT3/ITD negative

No SCT, %*
Donor, %
RHR
No SCT, %*
Donor, %
RHR
RR   89   50   5.9   50   51   1.0  
DFS   8   34   2.3   48   41   0.8  
OS
 
8
 
43
 
3.0
 
57
 
49
 
0.8
 


FLT3/ITD positive

FLT3/ITD negative

No SCT, %*
Donor, %
RHR
No SCT, %*
Donor, %
RHR
RR   89   50   5.9   50   51   1.0  
DFS   8   34   2.3   48   41   0.8  
OS
 
8
 
43
 
3.0
 
57
 
49
 
0.8
 

Donor indicates family donor available; RHR, relative hazard ratio; RR, relapse risk; DFS, disease-free survival; OS, overall survival.

*

Data from Table 4 of Gale et al.1

Data from Table 5 of Gale et al.1

Also, an “intention-to-treat” analysis, although statistically more “pure,” may not be clinically informative in this setting, as only 35 of the 68 FLT3/ITD-positive patients (51%) with donors actually received a bone marrow transplant (BMT) in first CR.1(Tab4) Clinicians are interested in whether FLT3/ITD-positive patients who reach the time of transplantation in first CR should receive a transplant. Thus, an “as-treated” analysis may be more informative, answering the question whether an allograft would provide an added benefit over chemotherapy-only in this population. Additionally, the comparison “no donor” group includes autograft recipients, where FLT3/ITD-positive patients had an improved survival over “no SCT” patients (48% vs 8%, P < .05; Gale et al1(Tab4)). Perhaps the most significant and informative data presented are that of relapse risk in the allograft recipients with and without FLT3/ITD.1(Fig4C) The figure clearly shows that allograft recipients had the same relapse risk regardless of the FLT3/ITD status, providing strong evidence that allogeneic stem cell transplantation negates the prognostic impact of FLT3/ITD. Surprisingly however, Gale et al argue against their own findings, suggesting that “this analysis may be subject to both bias, due to selection of patients for SCT and samples available for analysis, and the play of chance....”1(p366) Given that the authors have provided no evidence suggestive of selection bias or difference in sample availability, one wonders about the need to so strongly qualify the data.

The data seem to best suggest that while the FLT3/ITD is a poor prognostic indicator for AML patients receiving conventional chemotherapy, it does not seem to influence outcome after allogeneic SCT. Indeed, similar findings have recently been reported by others.2  This is similar to the situation of Philadelphia chromosome-positive acute lymphocytic leukemia (Ph+ ALL), a terrible prognostic factor with chemotherapy but not so with SCT.3  This issue is of considerable importance and some urgency. If allogeneic SCT indeed ameliorates the poor prognostic impact of the FLT3/ITD mutation, it should be studied as an important option in the care of patients with this genetic lesion and perhaps moved earlier in therapy, prior to impending relapse. This may be especially true if allograft can be provided with acceptable treatment-related mortality (TRM).

Allogeneic stem cell transplantation and FLT3/ITD status in AML

Meshinchi and colleagues disagree with our cautious interpretation of data that addresses the issue of whether or not allogeneic stem cell transplantation (SCT) abrogates the poor prognosis associated with a fms-like tyrosine kinase 3/internal tandem duplication (FLT3/ITD).1  They suggest that an “as-treated” analysis directly comparing allogeneic SCT and chemotherapy may be more informative than an “intention-to-treat” analysis, but we had presented this in Figure 3 of our paper. This shows that the impact of an FLT3/ITD on relapse is less in the allograft recipients than those who received an autograft or chemotherapy alone but that this difference is clearly nonsignificant. For there to be evidence that the increased risk of being FLT3/ITD positive was negated by allogeneic transplantation, there would need to be interaction between the allografted patients and the autografted or chemotherapy-only patients. There is no such interaction for either relapse or overall survival. We did not present the test for interaction over the 3 groups in Figure 3, but the result for relapse is χ22 = 1.8 (P = .4) and for overall survival it is χ22 = .9 (P = .6). Hence, there is no evidence at all of heterogeneity, precluding the conclusion that allogeneic SCT modifies the risk of FLT3/ITD positivity. Similarly, Figure 4C of our paper does not show that “allograft recipients had the same relapse risk regardless of the FLT3/ITD status” because of the small number of events and wide confidence intervals. The assertion to the contrary by Meshinchi et al is a confusion of a lack of evidence of effect with evidence of lack of effect.2 

To address the issue of whether allogeneic SCT modifies the poor prognosis associated with an FLT3/ITD, an “intention-to-treat” analysis in the form of a donor-versus-no donor comparison is preferable. This confirms the lower relapse rate associated with allogeneic SCT but indicates that this occurs in both the FLT3/ITD-positive and FLT3/ITD-negative patients. In their letter, Meshinchi et al construct a single table from our donor-versus-no donor analysis and the randomization to receive or not receive an autologous SC transplant. This is an injudicious comparison because patients with a donor are likely to have different characteristics from those entered into a randomization. The dangers of making such a comparison have been previously outlined.3  Furthermore, as we had pointed out in “Discussion,”1  the patients not receiving an autograft in the randomized comparison fared considerably worse than other patients in the trial who did not participate in the randomization and did not receive an autograft.

Meshinchi et al contend that “allograft may improve outcome in FLT3/ITD-positive AML” and we agree with this, at least as far as relapse is concerned. We are not saying that FLT3/ITD-positive patients should not receive an allograft, rather that the presence of an FLT3/ITD should not be factored into the decision whether or not to perform allogeneic SCT, as the benefit of the allograft is also seen in the FLT3/ITD-negative patients. We are also not excluding the possibility that allogeneic SCT may be of greater benefit in the FLT3/ITD-positive patients, but the data do not yet support such a hypothesis. It is an important issue, which is why we advocate larger studies or a meta-analysis. In the interim, statistical “purity” should not be abandoned.

Correspondence: Rosemary E. Gale, Department of Haematology, Royal Free and University College Medical School, 98 Chenies Mews, London WCIE 6 HX, United Kingdom; e-mail: rosemary.gale@ucl.ac.uk.

1
Gale RE, Hills R, Kottaridis PD, et al. No evidence that FLT3 status should be considered as an indicator for transplantation in acute myeloid leukaemia (AML): an analysis of 1135 patients, excluding promyelocytic leukaemia, from the UK MRC AML10 and 12 trials.
Blood
.
2005
;
106
:
3658
-3665.
2
Alderson P. Absence of evidence is not evidence of absence.
BMJ
.
2004
;
328
:
476
-477.
3
Wheatley K. Current controversies: which patients with acute myeloid leukaemia should receive a bone marrow transplantation?—a statistician's view.
Br J Haematol
.
2002
;
118
:
351
-356.
1
Gale RE, Hills R, Kottaridis PD, et al. No evidence that FLT3 status should be considered as an indicator for transplantation in acute myeloid leukemia (AML): an analysis of 1135 patients, excluding acute promyelocytic leukemia, from the UK MRC AML10 and 12 trials.
Blood
.
2005
;
106
:
3658
-3665.
2
Schlenk R, Krauter J, Fröhling S, et al. Postremission therapy with an allogeneic transplantation from an HLA-matched family donor seems to overcome the negative prognostic impact of FLT3-ITD in younger patients with acute myeloid leukemia exhibiting a normal karyotype [abstract].
Blood
.
2005
;
106
:
662a
. Abstract 2353.
3
Arico M, Valsecchi MG, Camitta B, et al. Outcome of treatment in children with Philadelphia chromosome-positive acute lymphoblastic leukemia.
N Engl J Med
.
2000
;
342
:
998
-1006.