Pooled Analysis Shows Predictive Features Linked With Variable Timing of Response to Venetoclax in AML

Courtney D. DiNardo, MD

Response to venetoclax (Venclexta)-based treatments are not always immediate in older patients with newly diagnosed acute myeloid leukemia (AML), explained Courtney DiNardo, MD, MSCE, who added that patient characteristics can be used to determine whether to continue or discontinue treatment.

To assess the likelihood of response to venetoclax-based therapy, investigators conducted a pooled analysis of 2 open-label trials that evaluated the BCL-2 inhibitor in combination with hypomethylating agents (HMAs; NCT02203773) and low-dose cytarabine (NCT02287233) in older patients with newly diagnosed AML.

Patients were split into 3 cohorts based on whether they achieved a complete response (CR) or CR with incomplete blood count recovery (CRi) within 2 cycles of therapy (early responders), after 2 cycles of therapy (late responders), or at no point during therapy (non-responders).

More than one-third of patients (35%) who achieved a CR/CRi required up to 7 cycles of therapy to achieve response. Of 197 patients, 42% (n = 83) achieved a CR/CRi within 2 cycles of therapy, 22% (n = 44) achieved a CR/CRi after 2 cycles of therapy, and 36% (n = 70) never achieved a CR/CRi. The median duration of response was 21.2 months (95% CI, 14.1—not reached [NR]) versus 8.1 months (95% CI, 5.3-14.9) in early and late responders, respectively.

Patients with a baseline IDH1/2 or NPM1 mutation were more likely to have a CR/CRi within 2 cycles of treatment, whereas patients with secondary AML who had not achieved a response by the end of the second cycle were more likely to never achieve a CR/CRi. Of 44 patients who achieved a CR/CRi after 2 cycles of therapy, 43% (n = 19) achieved a morphologic leukemia-free state (MLFS) within the first 2 cycles of therapy. Of 70 patients who never achieved a CR/CRi, 17% (n = 12) achieved a MLFS within 2 cycles of treatment.

Although these characteristics were predictive of the timing of response to venetoclax, they generally did not impact overall survival (OS), explained DiNardo. The estimated OS rates were comparable in early and late responders, respectively (1-year, 71% versus 75%; 2-year, 53% versus 42%). However, the median OS was not reached in early responders (16.9—NR) versus 18.0 months (13.3-27.8) in late responders.

"If I have newly diagnosed patient with de novo disease who is tolerating treatment well, even if they haven't had a response within the first 2 cycles, I'm going to keep them on treatment, especially if they've had leukemic loss reduction during the first 1 or 2 cycles, because they still have a good chance of going into remission later on," said DiNardo.

In an interview with OncLive, DiNardo, clinical researcher in the Department of Leukemia of the Division of Cancer Medicine at The University of Texas MD Anderson Cancer Center, discussed the results of the analysis evaluating the timing of response to venetoclax-based combinations, and related disease features, in older patients with AML.

OncLive: Could you discuss the current utility of venetoclax in AML?

DiNardo: In the United States, venetoclax is now considered a standard of care for older patients with newly diagnosed AML who are ineligible or otherwise inappropriate for intensive chemotherapy. In general, these are patients who are 65 years of age and older who previously would have received azacitidine or decitabine alone. In other countries, low-dose cytarabine is used frequently in that population as well. In the newly diagnosed setting, venetoclax is approved for use in combination with low-dose cytarabine, or azacitidine or decitabine.

What data were included in the analysis, and how were the response kinetics to venetoclax measured?

In our analysis, we combined data from 2 previously presented studies to identify the kinetics of response to venetoclax. We took patients who had been on the initial phase 1/2 study with an HMA and venetoclax, as well as patients on the phase 1/2 study with low-dose cytarabine and venetoclax and pooled them together.

There were 115 patients on the HMA study who received 400 [mg] of venetoclax, which was the [recommended] phase 2 dose. There were 82 patients on the low-dose cytarabine study, so we included 197 patients total.

We wanted to know when we can expect a response with venetoclax and how long we should continue this type of therapy if we don't see a response. When we're using azacitidine or decitabine alone, we know from decades of experience that it usually takes 4 to 6 cycles for patients to respond. We're very used to treating patients with continuous doses of azacitidine or decitabine without any remission. We know now from these phase 1 studies that responses with the combination happen really quickly, with an average of occurring after 1 or 2 cycles.

Therefore, if a patient has had 2 cycles of treatment, and they haven't gone into remission, should you discontinue treatment at that point? Or is there still some benefit to continuing the combination? Can patients respond after 2 cycles?

We grouped patients into 3 different groups: patients who had a CR/CRi within the first 2 cycles, any point after 2 cycles, or no response.

In the 66% of patients who did respond, we looked at response kinetics. Sixty-five percent of responses happened within the first 2 cycles, meaning that one-third of patients responded after 2 cycles of treatment. The majority of patients (95%) responded within 4 cycles of receiving an HMA with venetoclax. Approximately 86% of patients responded within 4 cycles of receiving low-dose cytarabine with venetoclax.

We looked at the characteristics of those patients who responded later to try to identify who those patients were. That way, a physician can tell whether a patient who hasn't responded by the second cycle of therapy [is likely to respond or not].

What did the results show?

The 2 most important things that were identified were first, patients who had leukemic blast reduction [were more likely to respond] within the first 2 cycles—particularly, those obtaining MLFS. About 40% of patients who responded after 2 cycles achieved an MLFS earlier within the first 2 cycles. Second, the type of AML, whether it was de novo or secondary, [also impacted the likelihood of response]. Patients with secondary AML were less likely to have a response after 2 cycles of treatment.

We used a Classification And Regression Tree model, which looked at these different characteristics. In general, patients who had an MLFS or newly diagnosed de novo disease still had a very reasonable likelihood of responding in subsequent cycles of treatment.

The other question was whether patients who responded later had responses that were not as long lasting or had [unfavorable] survival. In this study, the median OS in both groups, no matter when patients had their response was quite impressive. The OS was 18 months or more in both arms, with a 1-year estimated OS rate of about 75%, and an estimated 2-year OS of about 50%, suggesting that if patients were able to obtain a remission after the third or fourth cycle of treatment, it was still a very important end point.

What ongoing research with venetoclax can we expect to see?

The VIALE-A trial is the randomized confirmatory phase 3 study evaluating azacitidine with or without venetoclax. Those data are not quite ready to be presented yet, but hopefully will be soon. A press release announced that this was a positive study, so we're eagerly looking forward to being able to share those data in the coming month.

Jonas BA, Dinardo CD, Pratz K, et al. Timing of response to venetoclax combination treatment in older patients with acute myeloid leukemia. J Clin Oncol. 2020;38(suppl 15; abstr 7531). doi:10.1200/JCO.2020.38.15_suppl.7531