Different Subtypes of Acute Myeloid Leukemia Create Both Clinical and Research Challenges

A Conversation With Elihu Estey, MD

Get Permission

Elihu Estey, MD

Elihu Estey, MD

Acute myeloid leukemia (AML) is the most common leukemia in adults. Each year, about 20,000 Americans will be diagnosed with AML, and roughly 10,000 people in this country will die of the disease. AML progresses quickly, and unless treatment begins soon and is effective , the prognosis is grim. Only about one-quarter of treated patients are alive 5 years after diagnosis. Although typically treated uniformly, AML is not a single disease. Recognition of this is changing the way the disease is treated.

The ASCO Post recently spoke with renowned AML expert Elihu Estey, MD, Professor of Medicine, Division of Hematology at the University of Washington School of Medicine in Seattle, about current developments in the treatment of AML, among other issues. Dr. Estey is also Director of AML Clinical Research at Fred Hutchinson Cancer Research Center (Fred Hutch), with considerable experience in designing, conducting, and interpreting clinical trials in AML.

From Houston to Seattle

Please tell the readers a bit about your background and current position.

I spent 28 years at the MD Anderson Cancer Center, where I coordinated clinical research in AML and was heavily influenced by mentors such as Drs. Emil J Freireich and Michael Keating. I left MD Anderson in late 2007, largely because my wife had retired and preferred not to live in Houston anymore. I came to the University of Washington/Fred Hutch, where my responsibilities are very similar to those in Houston. Because I am old; have pretty much seen only AML; and been in two large AML centers, I suspect over the past 39 years, I’ve probably seen as many, if not more, patients with AML as anyone in the world.

Complicated Clinical Picture

Can you give us a glimpse into the complicated clinical picture of AML?

The best analogy is pneumonia. If you do a chest x-ray on a patient with a fever and see densities in the lungs, you’ll make a diagnosis of pneumonia. But there are many types of pneumonia. So how do you treat it? If you gave all patients with pneumonia penicillin, the response rate would be only about 5%. But if the pneumonia were caused by Streptococcus pneumoniae, the response rate would be 100%. And it’s similar with AML. Under the microscope, just as the lung densities in pneumonia patients look the same, so more or less do the blasts in the bone marrow in patients with AML. So what probably will happen in the future—although we’re a long way from there now—is that the various subtypes of AML will each be treated differently, much as is pneumonia. These therapies will be directed at the abnormalities specific to various types of AML: so-called “targeted therapies.”

A current problem with targeted therapies can be illustrated with a drug like Gleevec [imatinib]. As a single agent, it’s a wonderful therapy for chronic myeloid leukemia (CML). It probably cures many people. But if you give [imatinib] in the blast phase of CML, it’s not particularly effective. The patient will typically need a stem cell transplant on top of the drug. And the blast phase of CML is really analogous to other types of poor-prognosis AML.

So if a drug like imatinib works in the chronic phase but not in the acute phase, why would we expect a single targeted drug to work in AML? That pretty much defines where we are today.

"I think immunotherapy offers the most promising path forward, because, as currently understood, it might be more broadly effective than targeted therapies."
— Elihu Estey, MD

Tweet this quote

Combination Therapies

Given your comments, do you think targeted therapies offer promise in AML?

In the 1990s, people were beginning to say that chemotherapy had been usurped by targeted therapies. It was a message a lot of people wanted to embrace, because chemotherapy conjures up so many bad images. Of course, chemotherapy is itself targeted; otherwise there would be no remissions. And clearly chemotherapy is more tolerable today than 20 years ago due to advances in supportive care. And it appears targeted therapies become much more successful when combined with chemotherapy.

The most recent example is midostaurin (also known as PKC412), which is directed against FLT3. When you use it as a single agent, there is a response—yet no pronounced, obvious effect on survival. But when midostaurin is combined with chemotherapy, the results are far better, with people receiving the combination living longer than those given chemotherapy alone and (almost certainly) midostaurin alone. Recognizing the above, why don’t we have more combinations of targeted therapies, and why don’t we combine targeted therapy with chemotherapy earlier than we do now?

I believe the reasons are as much cultural and commercial as medical. Many drug companies are not necessarily eager to work in a collaborative way with other companies. If one company has targeted drug A and another company has targeted drug B, they may be in competition to bring their own drug to market, not necessarily to bring the best treatment to market.

There’s also a culture that so shies away from the potential of drug toxicity that progress is potentially hampered. If in an AML trial many people, as is often the case, die consequent to failure to respond that’s expected. But if a much smaller number die of toxicity that’s much less expected, even if these patients would have died a few months later of AML, noting the great majority of people who die of AML die with disease present than without.   

Nonetheless the unwillingness to accept even a modicum of toxicity can make a drug look bad. It is as if we are prisoners of the old dictum “First do no harm,” although with AML, it is plausible that first do harm may imply “then do no good.” This principle underlies the use of allogeneic hematopoietic cell transplant, which is probably the most reliable means to cure AML.

Challenges in AML Clinical Trials

As you’ve pointed out, AML is actually many diseases. How does this heterogeneity affect research in clinical trials?

Let’s say you put 400 patients on a trial of a single drug. But it’s evident that these 400 patients constitute 10 different disease groups, which need 10 different therapies. The challenges then become to have enough data from these smaller cohorts to maintain acceptable false-negative and false-positive rates. To make real progress, we need to run trials using combinations of targeted therapies and chemotherapies that are not too limited by toxicity-averse cultural bias.

Given the obstacles that you’ve outlined, where are we now in the treatment of AML?

Oftentimes, the best treatment options available in AML are those that are being studied in clinical trials; unfortunately, for a number of reasons, we don’t get many of our patients on trials.

Right now, the most common treatments in older patients with AML are decitabine and azacitidine. Bear in mind, these drugs while statistically better than their  predecessors often add  only a few more months of survival to patients who without AML would be expected to live another 10-15 years, even if they are 70 years old. Why not put more of these patients on trials?  I’m not sure why, although we do recommend trials for people predicted to do poorly with standard therapy. However, as evidenced by the data presented at the ASH [American Society of Hematology] over the past decade or so, most clinical trials, although often initially viewed as promising, are ultimately not successful. That’s a fact we need to recognize.

However, there’s no reason for pessimism, because we never know when a new agent will break through with remarkable results. We just have to keep moving forward with a large number of clinical trials, because it is difficult to predict in advance which will be best. If prediction were easier, I doubt there would be so many AML trials listed at

There are several examples of trials that have worked and produced excellent results. One is the trial on all trans-retinoic acid (ATRA) in acute promyelocytic leukemia, which was first tried in China in the 1980s. When the results were first announced, there was a lot of skepticism, because at that time China was considered a backward country. Then it was reproduced in the West, and we found out how it worked. So the takeaway is that we can’t always predict how a trial will work.

What is the most important step moving forward in AML research? 

We simply don’t know enough about the AML cell and its counterpart, the normal cell. Our major blockade to a cure is lack of knowledge. In generations to come, researchers who have that knowledge will look at us in the same way we look at the doctors who put leaches on George Washington, although they were perhaps the best doctors of their day. Offhand, I think immunotherapy offers the most promising path forward because it might be more broadly effective than targeted therapies as currently understood. ■

Disclosure: Dr. Estey reported no potential conflicts of interest.