The Aza/Ven reshuffle

In this issue of Blood,¹ Döhner and colleagues present a reanalysis of the study data that resulted in the regulatory approval of the Bcl-2 inhibitor venetoclax in combination with low-intensity chemotherapy (azacytidine, decitabine, or low-dose cytarabine). Mostly focusing on the most widely used version of this therapy, azacytidine + venetoclax (Aza/Ven), this new work focuses on a comprehensive molecular analysis of the acute myeloid leukemia (AML) specimens collected at study entry. The regulatory approval of these low-intensity venetoclax-based regimens initiated a dramatic global transformation in the treatment of AML,² offering an effective alternative to intensive chemotherapy for older patients.

The authors first used the molecular data to classify the study participants according to the 2022 European LeukemiaNet (ELN) risk stratification system.³ Given the fact that the ELN system is derived primarily from younger patients treated with intensive chemotherapy, it was perhaps not surprising that this classification was not particularly informative. Roughly 3 out of 4 participants fell into the ELN adverse-risk category, and responses and survival were improved by Aza/Ven compared with azacytidine + placebo across all 3 ELN risk groups. Döhner and colleagues developed a new prognostic model using the mutational status of just 4 genes, TP53, FLT3, NRAS, and KRAS, and from that derived 3 distinctly new groups. The adverse group consists of those with a TP53 mutation (irrespective of the variant allele frequency), the intermediate group consists of those with a signaling mutation (FLT3, NRAS, or KRAS), and the favorable group consists of everyone else.

Conveniently, also in this issue of Blood is a summary of the 2024 ELN risk classification for adults with AML receiving less intensive therapies,⁴ incorporating this new prognostic model. The new risk classification system also incorporates the recent findings that azacytidine in combination with ivosidenib is effective for IDH1-mutated AML⁵ and the fact that DDX41 mutations seem to confer favorable risk regardless of the type of therapy.⁶ 

In some ways, these findings and this new risk classification system codify what many already knew. There have been prior reports regarding the impact of these mutations in the outcome of patients treated with the nonintensive venetoclax combinations,⁷ and the 4-gene predictive signature was also identified in a recently published single-center study,⁸ providing important external validation. Without doubt, guidelines such as these are important for helping clinicians navigate through different treatment options. However, these guidelines also serve a broader purpose, with each of the 3 new risk categories laying the groundwork for research.

First, the adverse-risk group, comprised of patients with TP53 mutations, had a median survival of just over 5 months and appeared to derive no benefit whatsoever from Aza/Ven. It is already well established that they do not benefit from intensive therapy.⁹ Therefore this group should be enrolled in clinical trials specifically designed for them. Although to a degree this has already started happening, a further step could be taken: they could be treated as having a separate disease entirely (eg, “TP53-mutated myeloid neoplasm”), and any clinical trial enrolling such patients should be appropriately stratified, treating them as a separate category. In a randomized trial of any new therapy, if these patients are simply added to the mix, they will likely worsen survival in both arms, minimizing our ability to identify effective therapies for the patients with AML in the other 2 categories.

For the intermediate-risk category of patients with signaling mutations, clinical trials are already underway, at least for FLT3-mutated AML. The seemingly simple addition of a FLT3 inhibitor to Aza/Ven may be the answer for these patients but may not be, due to the myelosuppression induced by this triplet.¹⁰ Regardless of how we incorporate FLT3 inhibition into the less intensive regimens, we have clear avenues of exploration, and these are ongoing.

For the favorable-risk category of patients in this new system, things are less immediately clear, but still full of promise. Clinicians, never very fond of 7 + 3 for patients more than 60 years old, are becoming increasingly inventive in finding ways to declare someone unfit for intensive therapy (Grade 2 periodontal disease! Heel spurs!), and more and more patients under age 75 are being treated with less intensive venetoclax-based induction. At my own institution, two-thirds of patients newly diagnosed with AML between the ages of 60 and 74 are treated with a less intensive induction, mostly Aza/Ven. More than 3 out of 4 favorable-risk patients achieve a response to Aza/Ven, and not surprisingly, their fitness improves as well. As such, they become candidates for allogeneic transplant. Here, however, is an opportunity. Many of these responders, such as those with NPM1 mutations, will have very durable responses, potentially equaling or even outdoing the benefits of transplant. Measurable residual disease (MRD) could be used as a means to sort out those who might benefit from transplant. Currently this is routinely available for NPM1-mutated AML,¹¹ but this highlights the importance of developing and implementing MRD detection methods for a broader array of driver mutations.

With these important findings emerging from Viale-A and other studies, and with this reshuffling of risk groups, we have a clearer picture of where we are and a better idea of where we need to go to advance the field of AML therapy.

Conflict-of-interest: M.L. declares consulting or advisory roles for Daiichi-Sankyo, Amgen, Astellas Pharma, Bristol Myers Squibb, AbbVie/Genentech, GlaxoSmithKline, Syndax, and Takeda and expert testimony for Novartis.