Elihu Estey, MD

Acute Myeloid Leukemia Data From the ASCO20 Virtual Scientific Program

Question

Among the studies presented during the ASCO20 Virtual Scientific Program, which of those featuring novel therapies or treatment approaches in acute myeloid leukemia (AML) intrigued you the most?

Answer

In AML, a drug that receives a lot of attention is venetoclax. One of the studies that I found of interest was an update of a study of low-dose cytarabine plus or minus venetoclax.¹ I did a poster discussion of this abstract. Because venetoclax has received so much attention, I think these updates are useful.

In this particular study, the original paper referred to a difference in survival in favor of venetoclax at 12 months, but it was not necessarily statistically significant. They decided to update the results at 18 months and then it was statistically significant.

My interest in the paper is not necessarily that the results were now statistically significantly, but the fact that they had 6 additional months of follow-up. The amount of follow-up is important in AML. It has been shown that with enough follow-up in AML you can say to a patient that they are likely to be cured. Although I’m not sure that any patient is ever cured, their chance of getting AML even 10 years out is still higher than someone on the street, but once you get to a certain point the likelihood of relapse is low. You can tell the patient to plan the rest of their life. That time is somewhere around 3 years.

I have learned not to put too much faith in these abstracts, particularly those that report “encouraging” or “promising” results. Many years ago we looked at [the American Society of Hematology] abstracts of early phase studies in AML years back, and found that 85% had results that were “promising” or “encouraging.” Thirty-seven separate drugs had at least 1 positive abstract. With a minimum follow-up of 5 years, only 1 drug had been approved by the U.S. Food and Drug Administration in AML. Of course drugs today are different than then, and the criteria for drug approval may have changed. Nonetheless, 1 out of 37 is pretty low. That is why I look at many of these skeptically.

Question

In a study of the IDH1 inhibitor ivosidenib combined with venetoclax (+/- azacitidine) in IDH1-mutated myeloid malignancies, some patients experienced a composite complete response with minimal residual disease, while others experienced progressive disease in only 3 months.² How might learning the biomarkers of response help investigators pinpoint the groups that would stand to benefit the most from this treatment approach? Why or why not?

Answer

This is relevant to the discussion about drug approvals in the previous question. Composite complete response includes complete remission (CR) with incomplete blood count recovery (CRi), CR with incomplete platelet recovery (CRp) and is now beginning to include something called CR with partial hematological recovery (CRh). What these responses have in common is that criteria for them are less stringent than the criteria for CR. How did CR get recognized as an endpoint? It was shown that people who got CR rather than other responses lived longer than people who did not with the difference due to the time spent in CR. Ten years ago I was part of a paper using data from SWOG, EGOG, and the University of Texas MD Anderson that showed CRp was associated with shorter subsequent survival and relapse-free survival than was CR although CRp was associated with better survival than being resistant to therapy.

The survival benefit of CRh relative to CR remains unknown. Our patients do not care what we call a response. They only care if because of that response they are going to live longer. It is known that if you have minimal residual disease (MRD) your survival will probably be shorter. One thing we do know about CRi or CRp is they are much more likely associated with MRD than CR, suggesting CRi or CRp is less valuable than CR. We do not know if the same is true for CRh. I think the burden of proof is on those who focus primarily on composite CR (CRc) and disregard that the several components of CRc are not of equal value to patients.

Would identifying biomarkers help? Yes, but ‘biomarkers’ and the concept of ‘targeted therapy’ is another oversimplification. I don’t think it will be as simple as adding azacitidine to ivosidenib. We are exaggerating what we know. We may know how to treat 1 mutation, but another will arise. Identifying biomarkers will, in principle, help, but the ultimate test is whether identification will improve clinical outcome. What we know compared to what we need to know is very small.

Question

Results of a phase 3 study found that the addition of enasidenib to azacitidine in patients with newly diagnosed AML and mutated IDH2 significantly improved response rates and durations.³ Do you think the benefits seen were likely due to the upfront combination of enasidenib plus azacitidine, or as a result of the sequence of these treatments for those who crossed over to receive enasidenib after azacitidine (21% of patients)?

Answer

That is a good question that will likely require longer-follow up. If the results showed that patients receiving them together at the start seemed better than adding enasidenib, than, yes, it is better. They probably shouldn’t be as optimistic about these results as they are.

Question

Should palliative care be considered a “new standard of care” for certain patients with AML undergoing intensive induction? What is currently the standard of care for this patient subgroup?

Answer

Palliative care can have two meanings. The first is that the patient is not going to be treated. The second is that the patient will be treated, but will also have a palliative care consultation. No one can object to the latter! I have problems with the former since I think a very important way to reduce symptoms is to produce a remission. This will not happen without treatment. If standard therapies are highly unlikely to be successful I favor enrolling patients on a clinical trial. The fewer patients entered on these the slower our progress will be.

Question

Researchers reported there were complete responses seen in a phase 1 trial of an investigational bispecific T-cell engager (BiTE) product.⁴ These products typically require continuous infusion. Do you think this method of drug delivery could be a potential barrier to the adoption of BiTEs in AML?

Answer

The answer to this question will depend on the trade-off for the patient. If you are in a category of patients who have relapsed AML, which in the past would mean you would die within 6 months, then I don’t see how continuous infusion would be a barrier if the patient thinks such an infusion might improve their survival. In addition, now we have outpatient pumps that can provide continuous infusions, making these totally plausible in the outpatient setting.