Acute myeloid leukemia (AML) afflicts older individuals (median age 68 years), the majority of whom are unfit to receive intensive induction chemotherapy. In November 2018, the FDA granted accelerated approval to venetoclax (Ven), a small molecule selective inhibitor of the anti-apoptotic protein B-cell leukemia/lymphoma 2 (bcl-2), in combination with hypomethylating agents (HMA) and established a new standard of care for newly diagnosed (ND) elderly or unfit patients with AML [1, 2]. In recent years, the ubiquitous use of Ven for the treatment of AML has brought to the fore questions related to Ven sensitivity/resistance, optimal dosing schedule and duration of therapy. In the current commentary, we share lessons learned from Ven use in routine practice including our insights on genetic predictors of response and Ven dosing strategies in AML.

Preclinical and clinical trial observations

Ven targets bcl-2 which is overexpressed in leukemic CD34+ progenitor cells [3]. In vitro and in vivo studies involving AML cell lines, patient samples and murine xenografts exhibited sensitivity to treatment with Ven in the low nanomolar range (0.001–1 µg/ml) [4]. Furthermore, pre-clinical studies have also demonstrated synergistic activity of Ven with azacitidine [5]. Based on these preclinical observations, a phase 2 study of Ven monotherapy (800 mg daily continuously) was conducted in relapsed/refractory AML which showed overall response in 6 of 32 (19%) patients with complete remission (CR) in 6% [6]. Given the modest efficacy of Ven monotherapy in AML, subsequent clinical trials focused on Ven in combination with HMA or low dose cytarabine. In a phase 1b clinical trial in ND-AML (n = 145), patients received Ven (400–1200 mg daily for 28 days) in combination with decitabine (20 mg/m2, days 1–5, intravenously) or azacitidine (75 mg/m2, days 1–7, intravenously or subcutaneously); CR or CR with incomplete hematologic recovery (CRi) was documented in 67% of patients (CR in 37%), which formed the basis for accelerated approval of Ven for ND-AML in November 2018 [7]. Ultimately, Ven was granted full FDA approval in October 2020, following the phase 3 VIALE-A study (n = 431), which compared Ven (400 mg daily for 28 days) plus azacitidine (n = 286) to placebo-azacitidine (n = 145); in the particular study, 66% of patients achieved CR/CRi (CR in 37%) and survival was significantly improved with Ven-azacitidine combination therapy (median; 14.7 vs 9.6 months) [1]. Long-term followup from the VIALE-A study (median 43.2 months), confirmed survival benefit with Ven combination therapy, particularly in patients harboring mutations (MUT) in IDH1/2 (median 19.9 months, IDH1 MUT 10.2, IDH2 MUT 27.5 months) and those achieving measurable residual disease (MRD) < 10−3 (median 34.2 months) [8, 9].

Striking a balance: efficacy vs tolerability

Several studies including clinical trials and real-world series have shown that two-thirds of ND-AML patients receiving Ven-HMA achieve CR or CRi, while the remaining one-third fail to respond, and over half of responders inevitably relapse [8, 10,11,12,13]. Moreover, there is growing recognition that response patterns and survival outcomes following Ven-HMA are highly heterogenous and associated with mutational profile [14]. Recently, a pooled analysis of the phase 3 VIALE-A and phase 1b trials, which included 279 patients treated with Ven-azacitidine, unveiled a 4-gene molecular prognostic signature (mPRS) for Ven-HMA response and survival, based on the presence or absence of TP53 MUT, KRAS MUT, NRAS MUT, and FLT3-ITD and categorized patients into higher (FLT3-ITD, KRAS, NRAS, TP53 wild type), intermediate (FLT3-ITD, KRASMUT, NRAS MUT, TP53 wild type) and lower (TP53 MUT) benefit groups (CR/CRi in 77%, 59%, and 49% of patients, respectively), with median overall survival of 26.5 months, 12.1 months, and 5.5 months, respectively [15].

By contrast, different sets of genetic predictors for response and survival were reported in a Mayo Clinic study of 400 patients with ND-AML treated with Ven (median dose 200 mg daily) in combination with decitabine (n = 265) or azacitidine (n = 148), outside the clinical trial setting [12]. In the particular study, 38% patients achieved CR, and 24% CRi, resulting in CR or CRi in 62% of patients. CR/CRi rate was found to be highest at 87% in patients harboring one or more favorable mutations (NPM1, IDH2, DDX41) and no unfavorable mutation (TP53, FLT3-ITD, RUNX1) and lowest at 44% in patients with at least one unfavorable mutation and no favorable mutation [12]. Genetic risk factors which negatively influenced survival included European Leukemia Net (ELN)-defined adverse karyotype [16], KMT2Ar, TP53 MUT, KRAS MUT, and IDH2 WT, based on which a prognostic model was proposed with low, intermediate, and high-risk groups; respective median transplant-censored survival (3-year survival rate) were “not reached” (67%), 19.1 (33%), and 7.1 months (0%) [12]. Compared to the VIALE-A mPRS, the performance of the Mayo genetic risk model was shown to be superior (AIC 179 vs. 195 and AUC 0.77 vs. 0.69) [12]. Furthermore, among treatment responders i.e., patients achieving CR/CRi, adverse karyotype, KMT2Ar, KRASMUT, IDH2WT predicted inferior survival, allowing for a complementary response-stratified risk model [12].

In regard to toxicity, common hematologic adverse events observed in the VIALE-A study included grade ≥3 thrombocytopenia (46%), neutropenia (43%), and febrile neutropenia (43%) [1, 8]. Moreover, toxicities led to treatment discontinuation in 30% and were fatal in 25% [8]. Notably, six patients died while in ongoing CR/CRi, including 4 from infections (all with grade 3/4 neutropenia), Ven treatment duration was reduced to 21 days in 2 patients in CR, while the remaining 2 in CRi received Ven for 28 days [8].

Optimal Ven schedule

The VIALE-A study protocol recommended Ven for 28 days during the first cycle, while allowing dose schedule modifications include a delay in the initiation of next cycle, reduction in the number of venetoclax dosing days per cycle, or a combination of both, following achievement of CR/CRi [1]. Accordingly, among responders who received ≥6 cycles, starting cycle 6, median Ven dose duration was 15–21 days in 69%, and 7% received Ven for less than 15 days [8]. Complication rates were high and included grade ≥3 thrombocytopenia (55%), infection (69%), and hemorrhage (11%) [8]. In a separate post-hoc analysis of VIALE-A, Ven dosing schedule modifications occurred in 83% of patients; delay (median duration 10 days) in the next cycle after achieving blast clearance occurred in 74% of responders in the Ven-azacitidine arm, of which 72% patients experienced grade 4 cytopenia during the cycle delay [17]. Although, 50% of CR/CRi responders in the Ven-azacitidine arm received G-CSF post-remission, the median time to neutrophil recovery was similar regardless of G-CSF use [18].

As outlined above, the majority of patients receiving Ven-HMA experience significant toxicities related to myelosuppression, resulting in implementation of an abbreviated Ven schedule. In the article that accompanies this commentary, Willekens et al., report comparative outcomes from 82 patients with ND-AML receiving azacitidine × 7 days plus Ven × 7 days (7 + 7 regimen) vs standard dose Ven-HMA in cycle 1 (n = 111 receiving ≥21 days Ven) [19]. Topline results included similar CR/CRi (72% each) and overall survival (11.2 vs 10.3 months), however, early mortality at 8 weeks (6% vs 16%) and platelet transfusion requirements were substantially lower with attenuated Ven dosing (62% vs 77%) [19]. Furthermore, Ven duration did not appear to influence overall survival in patients with TP53 MUT, FLT3-ITD MUT, KRAS MUT, NRAS MUT, however, patients in the mPRS higher benefit group (FLT3-ITD, KRAS, NRAS, TP53 wild type), had inferior survival with 7 + 7 regimen vs standard dose Ven-HMA (14.1 vs 32 months) in univariate analysis [19]. In subsequent multivariable analysis limited to the higher benefit group, Ven treatment duration did not appear to independently impact survival outcomes [19].

Similar findings were observed in a Mayo Clinic study which included 270 ND-AML patients treated with Ven 14 days (n = 40, 15%), 21 days (n = 41, 15%), and 28 days (n = 189, 70%) during cycle 1 [20]. CR/CRi rates were similar in patients receiving Ven for 14 (68%) vs 21 (66%) vs 28 days (62%); corresponding MRD negative rates were 65%, 72% vs 81%; relapse was documented in 33%, 41%, and 40%, respectively [20]. In the aforementioned study, patients with ELN adverse karyotype showed numerically higher CR/CRi rates with Ven for 21 days (60%), vs 28 days (44%), vs 14 days (22%) [20]. Ven treatment duration did not appear to influence response rates in patients harboring one or more unfavorable mutations (TP53, FLT-ITD, RUNX1) without favorable mutations (NPM1, IDH2, DDX41); furthermore, CR/CRi rates were similar in patients with multi-hit TP53 mutations (20% vs 20% vs 23% with Ven 14 vs 21 vs 28 days) [20]. On the other hand, a non-significantly higher rate of grade 3 or higher infections was documented in patients on Ven for 28 vs 21 vs 14 days (28% vs 18% vs 20%; p = 0.29) [20]. Notably, survival was similar in patients receiving Ven for 14, 21 vs 28 days, with respective median survival of 18.6, 21.3, and 13.2 months [20].

Table 1 reviews pertinent findings from retrospective studies comparing different Ven dose schedules; all except one study showed non-inferior CR/CRi rates and a trend towards lower infection rates with shorter duration of Ven [19,20,21,22,23,24]. Overall, these findings require confirmation through an ongoing phase 2 randomized study (NCT03013998), comparing the efficacy and toxicity of 28-day vs 14 day-Ven-azacitidine in patients with ND-AML. On a similar note, metronomic Ven-HMA dosing has recently shown favorable results; in a phase 2 study with a weekly low-dose regimen of decitabine (0.2 mg/kg subcutaneously) and Ven (400 mg weekly) in AML (n = 21), 90% of patients continued treatment without dose interruptions/delays, CR/CRi rate was 52% (47% MRD negative), and overall survival was 16.1 months, with no treatment-related mortality [25].

Table 1 Studies on Venetoclax dosing schedule in acute myeloid leukemia (AML): 7 vs 14 vs 21 vs 28 days.
Full size table

In summary, the possibility that a shorter course of Ven in ND-AML might be as efficacious and less myelotoxic than standard (28-day) dosing is enticing. Moreover, currently developed FLT3/IDH inhibitor/HMA triplet combination therapies utilize an abbreviated Ven schedule in order to minimize toxicity. Accordingly, in our practice we prefer 14-day Ven dosing in patients with suboptimal physiologic age or patients at high risk for infectious complications. Regardless, sensitivity to Ven is strongly associated with genomic profile and warrants an individualized dosing strategy. Reduced Ven duration (or metronomic Ven) may be ideal to help prevent resistance as continuous Ven administration might place selective pressure on cells which may then develop resistance mechanisms.