To Transplant or Not To Transplant in First Remission?

Navneet Majhail, MD
Director of the Blood & Marrow Transplant Program at Cleveland Clinic Taussig Cancer Institute in Cleveland, Ohio
Alan K. Burnett, MD
Head of the department of hematology at the Cardiff University School of Medicine in Cardiff, United Kingdom

Most patients with newly diagnosed acute myeloid leukemia (AML) will achieve a first complete remission with standard induction chemotherapy. Durable remission or cure can only be attained, though, with further treatment – either post-remission consolidation chemotherapy or hematopoietic cell transplant. Questions about the optimal timing of transplant, as well as which patients will benefit from the procedure, remain an area of debate.

In this edition of “Drawing First Blood,” ASH Clinical News has invited Navneet Majhail, MD, MS, and Alan K. Burnett, MD, to debate the question: Should intermediate-risk AML be transplanted in first complete remission? Dr. Majhail will argue for transplant and Dr. Burnett will argue against.

Navneet Majhail, MD: In the debate about whether or not to transplant in first complete remission (CR1) in intermediate-risk AML patients, we have to acknowledge that the definition of “intermediate risk” is a moving target. There are several different types of diseases within the broad category of intermediate-risk AML – depending on a patient’s cytogenetic factors, molecular markers, and other prognostic markers.

Alan K. Burnett, MD: This is certainly an evolving field, and actually defining “intermediate risk” is complex. There are few data showing that patients’ survival is made worse by transplant, so I think the debate, really, is about the optimal use of the procedure – selecting the appropriate candidates who will benefit from transplant and whether it should be used in CR1 or later if needed.

Dr. Majhail: I think there have been great improvements in transplantation technologies and risk prediction that make transplant in CR1 the best option for properly selected patients.

Traditionally, cytogenetic markers have been used to classify patients’ risk levels, and, with the introduction of molecular markers like NPM1 and FLT3, we are able to further refine intermediate-risk patients – especially those who have a normal karyotype. We also have emerging data about the role of the allelic ratio (the ratio of FLT3-ITD/FLT3 wild-type) in predicting a patient’s risk of relapse.

But, until those data are validated, I would say the presence of an FLT3-ITD mutation in patients with normal karyotype or intermediate-risk disease qualifies patients for transplant in CR1.

Dr. Burnett: Within the intermediate-risk category defined by cytogenetics, though, we have to remember that not all mutations or markers indicate a poor prognosis. FLT3 and NPM1 have always been the biggest actors. In the UK for example, we use a validated weighted score to further determine a patient’s risk of relapse. In addition to cytogenetics, the score takes into account a patient’s white blood cell count at presentation, the response to induction chemotherapy, and whether a patient has secondary or de novo disease at presentation. This score has allowed us to identify about 20 percent of intermediate-risk patients (by cytogenetics) with a worse-than-average prognosis where transplant should be used.

Seventy-five percent of the high-risk patients defined by this score had an FLT3-ITD mutation. To decide whether the remaining 25 percent of FLT3-mutated patients are candidates for transplant in CR1, we use the interplay between FLT3 and NPM1. If a patient has an FLT3 mutation without associated mutant NPM1, he or she is at higher risk of relapse and benefits from a myeloablative transplant. For patients with FLT3 and associated NPM1 mutations, however, we do not have definitive evidence that they benefit from – or, for that matter, are harmed by – transplant.

Additionally, mutations in the CCAAT/enhancer binding protein-alpha (CEBPA) gene are associated with a relatively favorable prognosis especially if the mutation is bi-alleilic, so, for the approximately five to 10 percent of patients who have CEBPA mutations, transplant can be avoided.

Dr. Majhail: I agree, there are definite cases where the evidence does not support transplant. Even in our practice, patients with normal karyotype who have no other abnormalities but are either NPM1-mutated or have the biallelic CEBPA mutation are not offered transplant in CR1. Having said that, as we discussed earlier, the whole definition of intermediate risk is evolving.

Newer molecular mutations, such as DNMT3A or IDH1 and IDH2, are giving us more information about patients with a normal karyotype and higher risk of relapse. As we learn more about these mutations and their prognostic role in AML, I believe they will also help us better identify candidates for upfront transplant versus consolidation chemotherapy.

Looking at the existing data, I think certain subgroups of patients you identified might still be candidates for an allogeneic transplant in CR1. In a meta-analysis of 24 prospective clinical trials, for instance, Koreth et al. clearly showed that, for intermediate-risk AML patients, transplant provided an advantage for both overall and relapse-free survival.1

Dr. Burnett: Well, like you said, transplantation in AML is a moving field, so it seems like that meta-analysis is more a review of the history of transplantation, where the analyses were done on a “donor versus no donor” basis. This sort of analysis is a surrogate for randomization, but not all donors were transplanted in CR1 or at all, and not all “no donor” patients did not receive transplant in CR1.

Now, while transplant was clearly beneficial to event-free survival, when we start looking at the overall survival, I think that benefit becomes less clear. Obviously, there is some risk associated with forgoing transplant at CR1 because the patient may not reach second complete remission (CR2). The literature points to several features as predictive of relapse risk, but we must be careful not to assume that presence of these features means a patient will automatically benefit from transplant.

For example, would a patient with the monosomal karyotype benefit from transplant? This karyotype is associated with a very poor prognosis, but does a transplant help? It is our only treatment option, so we will go ahead with transplant, but I’m not sure it does our patients any good.

Yes, there are very few data suggesting that intermediate-risk patients undergoing transplant will have shorter survival, but we do know that a transplant is tough on patients. It can leave them with chronic debilitating problems. Therefore, we also need to focus in on the patients who do not need a transplant.

In that vein, residual disease measurement is a very influential predictor of outcome of transplant. . The data tell us that, if a patient is negative for minimal residual disease (MRD) and undergoes transplant, he or she will likely have a very good outcome; if a patient is positive with MRD, he or she won’t. MRD is also particularly interesting because it allows you to tailor treatment to the individual patient – rather than to use information about average outcomes. I think this could be a very important step forward in defining who needs what in terms of further treatment.

Dr. Majhail: Now, speaking of the literature on this topic, I wanted to look at a 2013 paper in the Journal of Clinical Oncology that analyzed outcomes and survival in patients in the Medical Research Council AML10, 12, and 15 trials, which included 3,415 patients who entered CR1 (2,029 were intermediate-risk patients).2

Of the total population, one-third received a transplant in CR1. Overall, 1,271 patients (60%) who did not receive transplant in CR1 relapsed (61% of these patients were considered intermediate-risk according to their cytogenetic profile). Only half of these patients (642 people) eventually achieved CR2 (59% of whom were intermediate-risk patients).

Essentially, of all patients who relapsed without a transplant in CR1, only one-third went on to receive a transplant in CR2. The numbers were similar when considering the specific subgroup of intermediate-risk patients, in which, again, only one-third of all patients who relapsed received a transplant.

When patients were transplanted in CR2, their survival was lower than what might be expected for transplantation in CR1. Keeping in mind that these data represent patients enrolled on those trials, I can deduce that many intermediate-risk AML patients will eventually relapse and, among those who relapse, only half will eventually achieve CR2. At the end of the day, a majority of patients do not end up getting to a transplant and are not able to receive any of the procedure’s potential benefits. Even if you disagree with universal transplant in CR1 in these patients, you have to agree that there is value in at least considering it for appropriate patients.

Dr. Burnett: Our interpretations of those data are a bit different: You are correct that only about 50 percent of patients with intermediate risk will achieve CR2 and about one-third of the patients got a transplant in CR2.

Generally speaking, about 20 to 25 percent of patients can be rescued if there is a donor readily available. To me, these data tell me that a proportion of intermediate-risk patients can be salvaged. In an event-free analysis this group is excluded. These numbers have to be added to the numbers of patients who survive with chemotherapy in CR1. This illustrates why overall survival is an important endpoint to consider, rather than just event-free survival.

Dr. Majhail: One has to keep in mind that survival after transplant in general has improved considerably over the past few decades. The relatively high rates of transplant-related mortality that were reported in historical studies are really not reflective of the supportive care we routinely offer patients today. We are getting better at choosing patients for transplantation, including incorporating newer factors into the decision process.

Of course, the decision gets even more complex in older patients. AML is typically a disease of older patients, who carry more comorbidities than younger patients do. Newer instruments, like the HCT Comorbidity Index and the EBMT risk score, are helping us better classify the risk of transplant-related mortality. Also, availability of a donor is no longer a barrier to transplantation; with an increasing pool of unrelated donors and with umbilical cord blood and haplo-identical donor sources becoming more mainstream, essentially all patients who need an allogeneic transplant have a potential donor.

To change direction a bit, we should discuss older patients with AML – a group that has universally poor outcomes with chemotherapy only and should be considered, if appropriate, for transplant in CR1. Reduced-intensity conditioning (RIC) transplantation offers a relatively low risk of transplant-related mortality and opens up the possibility of allogeneic transplantation for older patients with AML. RIC transplant seems to produce lower relapse and better leukemia-free survival than chemotherapy alone. One can argue that transplant should be strongly considered for appropriately selected older patients with intermediate-risk AML in CR1.

Dr. Burnett: For an older patient who is in otherwise good condition, there is no question in my mind that RIC transplant should be considered. In our experiences with older patients – in about 100 to 150 patients older than 60 years – there is little evidence that these patients benefit from consolidation chemotherapy. The odds of any level of survival after relapse are low for these patients, so, obviously, delaying a transplant when the opportunity is there is inadvisable.

However, when our investigators looked into RIC transplants in older patients, they found that, in fact, less than 10 percent of older patients in remission end up going down this route. I do think a proportion of older patients could benefit from the RIC transplant but are not currently being offered it due to many factors – perhaps the comorbidity burden rules out this approach, or the donor might not be medically suitable.

People used to believe that RIC transplant would never work in AML, citing its poor graft-versus-leukemia effect and the fact that it takes time for RIC transplant to produce some effect – by which time the patient has probably relapsed. I think we are at a point now where we can control leukemia, and RIC transplant is a feasible way to do so. Ultimately, though, the choice will depend on what we learn in future years about the best conditioning schedule for RIC transplant.

Dr. Majhail: With all the new research in the field of AML, I think we can expect perspectives in this debate to continue to change. Newer therapies, including FLT3-inhibitors and IDH2 inhibitors, may change the “control arm” for transplant.

We can at least agree that all patients with intermediate-risk disease should get a transplant consult early in their disease course. The last thing we want is a delayed transplant in a transplant-eligible patient, or to see a patient relapse while waiting for transplant. So, treatment of these patients should be a collaborative effort between the leukemia physician and the transplant physician.

Dr. Burnett: I agree – the feasibility of transplant has to be known very early in a patient’s management. Perhaps the discussion we need to have with patients is not “to transplant or not to transplant,” but whether he or she needs the transplant now or later, after demonstrating an absolute need for it.


  1. Koreth J, Schlenk R, Kopecky KJ, et al. Allogeneic stem cell transplantation for acute myeloid leukemia in first complete remission: systematic review and meta-analysis of prospective clinical trials. JAMA. 2009;301:2349-61.
  2. Burnett AK, Goldstone A, Hills RK, et al. Curability of patients with acute myeloid leukemia who did not undergo transplantation in first remission. J Clin Oncol. 2013;31:1293-1301.
  3. Burnett AK, Hills RK, Wheatley K, et al. Sensitive risk score for directing treatment in younger patients with AML. Blood. 2006;108:10a (Abstract 18).