Dr Navarro: It's my pleasure to introduce Dr Fred Appelbaum. Dr Appelbaum is president and executive director of the Seattle Cancer Care Alliance. He'll be speaking today about transplantation in AML in the older adult.

Frederick R. Appelbaum, MD: I am planning to talk predominantly about the indications for transplantation in older individuals.

We all are aware of the indications for allogeneic transplantation in younger individuals, that is, patients who are younger than age 50. Numerous studies involving individuals with AML who were treated initially with chemotherapy, achieved a complete remission, and then went on to a transplant if they had a matched sibling or were treated with conventional chemotherapy if they didn't have given us a fairly firm understanding of who among the younger individuals should undergo transplantation.^[16-18] These include individuals who have failed primary induction, generally defined as people who have failed 2 cycles of standard 7+3 or 1 cycle of a high-dose cytarabine-containing preparative regimen. People who do not get a transplant who have failed primary induction inevitably die of their disease; with transplantation, although the cure rates are not great, perhaps 20% to 25% of people can still be salvaged.

For patients who do get a first remission and are younger than age 50, allogeneic transplantation has been shown to improve overall survival except in those who have favorable-risk disease. We define favorable risk as those who have CBF leukemias or those who have normal cytogenetics with a mutation in NPM1 or CEBPA but are FLT3 wild type.

Finally, we know that patients for whom initial chemotherapy fails -- that is, relapse after achieving a complete remission -- generally benefit from a transplant in virtually every category of disease, except perhaps those who have a very long first remission and who have favorable-risk cytogenetics. Because those individuals are incredibly rare, we generally say that transplantation is indicated among people younger than age 50 who have primary induction failure, who have achieved a first remission (except for those who have a favorable-risk disease), and for those with recurrent leukemia. That is what we know.

What we don't know, however, is what the indications are for allogeneic transplantation in patients who are over age 50. I'll go up to age 70 in my initial remarks, and then a little bit later look at those who are over age 70.

This became a big question when researchers started showing that you could transplant patients who were in their fifth, sixth, and even seventh decade of life if you used reduced-intensity-conditioning (RIC) regimens. A lot of this work should be credited to Rainer Storb and my colleagues in Seattle who showed that you could get grafts with much reduced intensity; Simon Slavin and others in Europe have also shown similar outcomes. We published results showing that patients who were over age 50 and had AML in first remission could achieve 5-year survival rates between 30% and 40% if they undergo transplantation in first remission.^[19] The results were very similar if they underwent transplantation in second remission.

Questions have always come up regarding how selected these patients were and what the effect of age on outcome really is. So people started to look at RIC more carefully, and in the time since these results were published many more studies have been published that allow us now to at least try to dissect some of this information.

This is a study from the Center for International Bone and Marrow Transplant Research (CIBMTR) from 2010 that looked at the overall outcome of transplantation for patients with AML or MDS, who were in their 40s to 50s, 50s to 60s, or 60 to 65.^[20] The study looked at a large number of patients who underwent transplantation between 1995 and 2005.

Both the nonrelapse mortality and relapse rates among patients in first complete remission were very similar, no matter which decade of life the patient was transplanted in. There was virtually no difference by age. Similarly, there were no significant differences by age with overall and disease-free survival. The only thing that predicted for worse outcome was patients who had greater human leukocyte antigen (HLA) disparity. About half of these patients were nonidentical, not matched siblings -- they were unrelated donors, some of whom were mismatched for a single antigen -- who had unfavorable cytogenetics or who had poor performance status. In multivariate analysis, low performance status, mismatched unrelated donor, and unfavorable-risk cytogenetics predicted for poor outcome.

Of course, these results do not say who should be transplanted because there was no comparison with chemotherapy. We also don't know how carefully these patients might have been selected. As a result, mostly in retrospective studies, people have tried to compare outcomes in patients who were over age 50 using chemotherapy with transplantation.

This is a study that was published by the Japanese group in 2011.^[21] It's a survey -- not a study from a single center. Taking a survey of 67 different centers throughout Japan, the researchers identified about 1000 individuals who had AML and who achieved a first complete remission. They then tried to compare the outcome of allogeneic transplantation vs chemotherapy.

The biggest flaw was that more than 600 patients never had their HLA type checked. Was it because they were so healthy, so young, and had such favorable-risk disease that they didn't bother to transplant? Or was it because they were so old and so sick that they couldn't be transplanted? One can't answer, so there's a huge potential for selection bias. In fact, only 15% of those who had AML in first remission actually ended up being transplanted in this comparative group.

When they looked at the outcomes, what they found was that the chance for relapse was much higher for those who had chemotherapy as opposed to transplantation. They also found that the relapse-free survival and overall survival were significantly better with transplantation than with chemotherapy. These do reach statistical significance, but again, there is such a heavy selection bias that it's very difficult to determine whether or not this proves the superiority of transplantation.

A study from the CIBMTR tried to do the same thing as the Japanese study.^[22] The investigators looked at AML patients in first remission ages 60 to 70 who between 1999 and 2005 received a reduced-intensity-conditioning transplant. Using the Cancer and Leukemia Group B (CALGB) database, they did a 1-to-1 comparison based on age and risk to try to compare outcomes, and the outcomes were similar to what the Japanese group had seen. That is, the chance for relapse was much higher with chemotherapy than with transplantation, and there was a nonsignificant improvement in overall survival with transplantation compared with chemotherapy.

What the group did here, which is probably an appropriate thing to do, was to not look at survival until 4 months after patients had achieved their first complete remission. This way they could eliminate any lead-time bias in the transplant group. So they only looked at people who were alive in complete remission 4 months after diagnosis -- which explains this peculiar curve that starts 4 months after the start of the trial.

Again, this is retrospective, comparative data. We do not know how important selection bias might be. And this figure shows the overall survival in that comparison between chemotherapy and transplantation.

The trend toward an advantage for transplantation also played out in every cytogenetic risk group, including favorable risk. I should note that there were only 10 patients in the favorable risk group, so that can be ignored -- although the outcomes are better, the numbers are so small that it really doesn't prove anything. But whether you had normal, intermediate, or adverse cytogenetics, overall survival at 3 years was better in the group that underwent transplantation compared with the chemotherapy group. Again, this was not a prospective, randomized study. It was a retrospective study looking at CALGB data and trying to match it up 1-to-1.

This was a study that Elihu Estey published in Blood in 2007 when he was at MD Anderson.^[23] Rich Stone and others have titled this study the "many-were-called-but-few-answer" study because they tried to do transplantation in older patients at their institution, but out of the 259 initial patients fewer than 10% actually underwent transplantation. This suggests there was a huge selection bias -- raising the following question: Although the group that underwent transplantion did better than those who received chemotherapy, was this difference really because they were inherently healthier and better patients?

We asked ourselves whether these results were due to peculiarities either at MD Anderson or in how well transplant was accepted at the time. We did a study that was presented at ASH 2011 looking at every patient with AML between the ages of 50 and 75 who was induced at our institution.^[24] We found that there were 81 such patients, and we attempted transplant for all of them. We succeeded in having 65% of them undergo transplantation, so not the 10% we saw at MD Anderson. So, through some effort, we got two-thirds of our patients between the ages of 50 and 75 to transplantation.

The reasons patients didn't go to transplant included early relapse (10%), comorbidities (7%), either the patient or the physician did not want to go forward because of the perceived risks (6%), insurance did not cover it (3%), and a smattering of other (9%). Among the reasons why patients did not go forward, lack of donor was never a reason. We were willing to transplant patients using matched siblings, matched unrelateds, single-antigen mismatched unrelateds, or cord blood. Here you can see the percentage of relateds, unrelateds, and cord blood transplants we performed.

Is it possible to do a prospective, comparative study of transplantation in patients between ages 50 and 75 to really test whether transplantation is better than chemotherapy? You can't do it on a donor-vs-no-donor basis -- or at least it would be very difficult unless you would be willing to take an arbitrary stand such as "we're only going to do fully matched unrelateds and we won't do single-antigen mismatched unrelateds," although data suggest this outcome would not be different. It would have to be essentially an arbitrary decision. So this becomes a very tough study to do. It also becomes very tough considering that in our institution, which I would say "smiles" on transplantation, we worked hard but were able to get only two-thirds of patients. That is still one third that won't get to a transplant. And what do you do with those individuals? Do you do the study on an intent-to-treat basis? If you do it on an intent-to-treat basis, you have to have a big, effective transplantation to make up for the third of patients who never get a transplant and you don't expect to get a transplant.

These data from the CIBMTR demonstrate why a decision to allow people to undergo transplantation with a matched unrelated donor, but not a single-antigen unrelated donor, would be a fairly arbitrary one.^[25] The study looked at 2223 patients who underwent transplantation for AML. The median age was 50, and half were transplanted in first remission and half beyond. Also, half received ablative therapy and half nonablative. At 2 years, survivor outcome was really the same whether they had a matched sibling, a matched unrelated donor, or even a single-antigen mismatched unrelated donor (although the early mortality over the first year is a little higher). But in terms of long-term survival, the outcomes are very similar, and this is not in a small number of individuals -- it is over 2000 individuals. So again, a donor-vs-no-donor study would have to be a fairly arbitrary study.

These data, again from the CIBMTR, looked at cord blood vs unrelated donor transplants for patients with AML who received RIC regimens.^[26] These individuals were between ages 50 and 70, and half had acute leukemia in first remission, half beyond. There was no statistically significant difference between an 8 of 8 peripheral blood, a 7 of 8 peripheral blood, and a double umbilical cord transplant in patients receiving the regimen that we were using in Seattle and that was being used in Minnesota at the time: total-body irradiation, cyclophosphamide, and fludarabine. There were other cord-blood regimens used, and whether this result was just chance or related to patient selection is difficult to say, but the other cord blood regimens did significantly worse than this one.

It also may be that since these data are mostly from Minnesota and Seattle, experience played a role vs results from some of the smaller centers. Again, it could be patient selection or chance. But with a regimen of total-body irradiation, cyclophosphamide, and fludarabine, there was not much difference in survival using double cords, even in patients ages 50 to 70, with reduced-intensity conditioning.

As of now, we don't know whether transplantation is better than chemotherapy in patients in first remission because there has not been a prospective randomized study. The retrospective comparative studies suggest an advantage to transplantation, but one has to take that for what it's worth. These are retrospective comparisons done as well as we can do them. It's going to be very hard to do a comparative study.

We do need to do comparative studies because things change with time. These are the outcomes by SWOG studies.^[27] From 1980 to 1990 to 2000, cure rates for AML continued to go up with no new drugs (we still have daunomycin and cytarabine), so the results are more likely based on us using our current drugs in a better way as well as better supportive care.

This abstract from ASH 2012 looked at the chance of dying following induction with chemotherapy for AML.^[28] We combined a large number of patients from SWOG and MD Anderson with a median age of about 60 years old. From 1991 to 1995, there was about an 18% chance of dying during the first 3 months after induction chemotherapy. In 1995 to 2000, it dropped to 14%. This all occurred with intensive induction therapy -- all 7+3. The figure fell to 9% and today it's less than 3%.

For those older than age 60, the chance of mortality during induction dropped from 22% in the 1990s to 15% and now to 5%. Again, it was always using 7+3 for AML. So these results are due to better supportive care. I think the antifungals are probably the biggest change. It may be that we're paying more attention to hospital-acquired infections and that we have a better handle on platelets, but I think most of these changes are due to antiviral and antifungal medications as well as better hygiene within our hospitals.

The same thing is true with transplantation: As chemotherapy has gotten safer and possibly better, so too has transplantation. This study, published in The New England Journal of Medicine in 2010,compared 1500 patients who underwent transplantation in Seattle on average in 1995 vs those a decade later on average in 2005.^[29] During that time, we halved the mortality rate associated with allogeneic transplantation. This was not due to reduced-intensity transplants. The mortality rate was halved for both intensive and reduced-intensity transplants. Again, it was better supportive care that led to these improved outcomes.

We have to figure out a way to continue to compare outcomes of chemotherapy with transplantation, although we don't think things are necessarily changing. I think they are in ways that may be subtle and unpredictable.

So currently, what do we recommend for older individuals -- those who are age 50 to 70 -- as far as transplantation is concerned? Based on the retrospective comparisons, I would recommend that we use the same indications that we use for patients who are younger than age 50. We don't have a prospective study at this time, and these outcomes must be discussed with the patient because there are going to be patients who are less healthy after transplant.

Turning now to patients for whom primary induction fails, this is a study that deserves some attention.^[30] I mentioned that transplantation, including for patients with unrelated donors, is an appropriate therapy for patients with primary induction failure who are under age 50. Even if you don't have a matched sibling, transplantation for patients with primary-induction failure who find an unrelated donor appears to give a reasonable outcome. In this study, about 25% of patients were alive after receiving an unrelated donor transplant for primary induction failure AML. Outcomes were little better if they were transplanted earlier and if they received just 2 courses of induction rather than if they had continued with more therapy.

Factors that were associated with improved outcome with transplantation for primary induction failure included a shorter time to transplant, having fewer than 3 induction courses (so just 2 induction courses), and having 38.5% or fewer blasts. Finally, about half of these patients received RIC. Although it wasn't significantly better, there was a trend toward improved survival with RIC.

Finally, what about patients for whom first-line therapy has failed? Approximately 40% of patients are alive and disease-free 3 years after transplantation for AML in second remission using RIC based on data in Seattle.^[19] So we generally recommend that patients who experience relapse be considered for transplantation.

Is every patient who experiences relapse a transplant candidate? Probably not. This is a study that we published with a group from Boston recently in Blood, that tried to determine disease risk based on whether patients had favorable, intermediate, or adverse cytogenetics, and whether they were in active relapse or remission.^[31]

Patients with adverse cytogenetics and in active relapse fell into the high/high group. Patients in the high/high group didn't do very well. So one would question whether transplantation is appropriate for that group; that is, patients with active relapse and high-risk cytogenetics. Everyone else had survival rates that were 20% or better, even patients who were in relapse sometime after transplant. Still, not everyone is a candidate, and I would say that transplantation might be viewed as futile for the high/high group.

What are the indications for allogeneic transplant for patients with AML who are ages 50 to 70? They are very similar to those for patients who are younger than age 50 -- that is, patients who experience primary induction failure, patients who are in first remission, except those who have favorable-risk cytogenetics, and anyone who has recurrent disease, except those who have very advanced recurrent disease with unfavorable cytogenetics.

Now of course, we've just been talking about age, and I want to make the point that age per se may not be so important. This is a study that we did with Roland Walter looking at patients with AML who were undergoing chemotherapy.^[32] We asked what the factors were that predicted for early mortality with chemotherapy in patients with AML. There were a group of factors: performance status, age, platelet count, creatinine, albumin, having secondary AML, and peripheral blasts when you had primary induction. We then found we could model these findings to predict who was more likely to die during induction therapy.

If you think of flipping the coin as 50-50 and perfect prediction as being 1, our model provided a 0.82 chance of predicting who was going to die vs not during induction therapy. When we removed age from our model, it still provided a 0.8 prediction of who would die during induction. The point therefore is that it isn't age per se, but these other factors – many, but not all, of which accompany age -- that we should pay attention to.

Mohammed Sorror at the Fred Hutchinson Cancer Research Center spent a lot of time studying this question, and he developed a comorbidity index that does not include age.^[33] The comorbidity index succeeded in predicting who would do well in patients who underwent transplantation in first remission. Patients with a comorbidity index of 0 had the best survival. As you can see, survival rates for those with a comorbidity index of 3 or greater fell precipitously. We verified these results using a second institution, MD Anderson. Scores of 0, 1 to 2, and greater than 3 on the comorbidity index had a decreasing association with survival.

The comorbidity index is not the same thing as performance status. There is some overlap, but they are not the same thing. This study of transplantation outcome looked at both the comorbidity index and performance status.^[34] Some patients who had a comorbidity index of 0 or 1 had a very good performance status whereas others had a poor performance status. Furthermore, some patients who had a higher comorbidity index had a good performance status, whereas others had a poor performance status. They are independent, and they both can be used to predict outcome.

Look down these columns. This study of more than 3000 patients with AML who underwent transplantation compared the comorbidity index and age to assess nonrelapse mortality and overall survival.^[35] Nonrelapse mortality goes up with increased comorbidity index across every age group, but the figures don't change as much when you look by age group.

Similarly for survival: Survival goes down with each age group as the comorbidity index goes up, but if you look by age, there is not much change. The point is that comorbidity index is a much more powerful predictor of outcome of transplantation for AML than is age. It predicts more for nonrelapse mortality, and it predicts more for overall survival.

Can patients who are over age 70 undergo transplantation? We have done this with a handful of patients -- 32 patients ages 70 to 75 with AML. Here you can see that there was still a 30% to 35% survival rate at 4 years.^[36] The curve gets unstable further out. There were 2 late deaths, 1 from graft-vs-host disease (GVHD) and 1 from an unrelated health problem. So it is possible to carry out transplantation in highly selected patients, even in this age group.

In sum, the indications for allogeneic transplant for patients with AML, ages 50 to 70, are similar to those for patients younger than age 50. Transplantation can be successfully conducted in selected patients older than age 70. And comorbidities are far more important than age in predicting outcomes of allogeneic transplantation.