The Bone Marrow Blasts on Day 14 during the Azacitidine Plus Venetoclax Regimen As the Induction Therapy Can Predict the Outcome in Unfit Patients with Newly Diagnosed Acute Myeloid Leukemia

Background 28-day venetoclax plus hypomethylating agents as a standard induction is recommended for patients with acute myeloid leukemia (AML) except APL who are unfit for intensive chemotherapy (IC). However, it is intolerable for a certain proportion of patients due to myelosuppression or other toxicity. Whether shortening the administration duration of venetoclax affects treatment responses at the end of the first induction remains unclear. Furthermore, it is rarely reported whether the early assessment using day-14 bone marrow (D14-BM) morphology predicts the outcome. This study aims to explore the predictive value of venetoclax administration duration and D14-BM blasts during azacitidine plus venetoclax therapy on treatment responses and survival in patients with newly diagnosed AML.

Methods Data from consecutive newly-diagnosed adults with AML unfit for IC receiving azacitidine and venetoclax as induction therapy at Peking University People's Hospital were reviewed. Bone marrow morphology was evaluated on day 14 and day 28 of the first induction and after each cycle of following treatment. Measurable residual disease (MRD) was assessed by flow cytometry after each cycle. After induction therapy, patients who achieved CR/CRi and were fit for IC received consolidation with intermediate/high-dose cytarabine based regimen or transplant; otherwise, they continued with azacitidine plus venetoclax until relapse. Patients not responding to 2 cycles of induction therapy or relapse switched to other salvage therapy. The optimal cut-off value of D14-BM blasts for predicting CR/CRi after the first cycle of induction was determined by receiver operating characteristic (ROC) curve method. Logistics regression or Cox regression analyses were used to explore variables associated with treatment response and survival.

Results Data from 247 consecutive patients with AML were included. 142 (58%) patients were male. Median age was 60 years (IQR, 48 - 67 years). 73 (30%) patients had ECOG PS > 2. 62 (25%) patients had ≥ 1 comorbidity(ies). According to the ELN 2022 recommendations, 76 (31%) patients were in favorable-risk; 48 (19%), intermediate-risk; and 123 (50%), adverse-risk. Median D14-BM blasts was 2.5% (IQR, 0.5 - 8%). The cut-off of D14-BM blasts was determined as 5% by ROC curve. D14-BM blasts < 5% was observed in 154 (62%) patients. 124 (51%) patients received venetoclax for less than 28 days during the first cycle of induction therapy. The most common reasons for venetoclax discontinuation were myelosuppression-related toxicities (n = 114), followed by heart failure or arrhythmia (n = 7) and other reasons (n = 3). The median administration duration of venetoclax during first induction was shorter in those with D14-BM blasts < 5% compared to those D14-BM blasts ≥ 5% (21 days vs. 28 days, P = 0.002).

A total of 174 (70%) and 123 (49%) achieved CR/CRi and MRD negativity at the end of first induction therapy, respectively; 204 (83%), CR/CRi ultimately. With a median follow-up of 9 months (IQR, 5 ‒ 18 months), 44 received transplant, 54 relapsed, and 59 died of no response, relapse or transplant related mortality. The 18-month survival rate was 68% (95% CI: 60, 76%).

D14-BM blasts < 5% resulted in a higher CR/CRi (92% vs. 36%, P < 0.001) and MRD negativity rates (67% vs. 22%, P < 0.001) at the end of the first induction, a higher final CR/CRi rate (96% vs. 60%, P < 0.001), and higher survival rate at 18 months (76% [68%, 85%] vs. 53% [38%, 67%], P < 0.001). In the multivariate analysis D14-BM blasts < 5% was significantly associated with higher CR/CRi and MRD negativity rates at the end of the first induction (OR = 19.3 [95% CI, 7.2, 51.5], P < 0.001 and OR = 7.1 [3.9, 13.0]; P < 0.001) and final CR/CRi rate (OR = 20.3 [9.9, 41.7]; P < 0.001), as well as favorable survival (HR = 0.3 [0.2, 0.5]; P < 0.001). However, venetoclax administration duration did not affect the treatment responses and survival. In addition, older age, ELN high risk, increasing LDH, and secondary AML had adverse impact on survival.

Conclusion Our retrospective study suggested that D14-BM blasts < 5% during azacitidine plus venetoclax induction therapy predicted better treatment responses and favorable survival in unfit patients with AML. Whether shorter administration duration of venetoclax in those with lower D14-BM blasts is feasible required further investigation.