AML-331: Real-Life Evidence of Treatment with Venetoclax in Patients with Relapsed or Refractory Acute Myeloid Leukemia

Context: The prognosis of patients with relapsed or refractory acute myeloid leukemia (RR-AML) is often poor, and treatment options are limited. The good results of venetoclax (VEN) in untreated AML have led to its off-label use in RR-AML. Objective: We aimed to analyze the effectiveness of the use of VEN in patients with RR-AML reported to the Spanish PETHEMA AML epidemiological registry. Design: We conducted a retrospective, multicenter, observational study. Patients or Other Participants: Patients with AML-RR who received treatment with VEN in the PETHEMA group of hospitals were included. Interventions: No intervention was performed. Main Outcomes Measures: Response was assessed as complete remission/complete remission with incomplete hematologic response (CR/CRi), partial remission (PR), and overall survival (OS). Results: Fifty-five patients were included, 33 men and 18 women, with a median age of 68 years (25–82). With a median follow-up of 167 days, 10/51 patients (19%) continued to receive VEN. Patients received a median of 2 cycles (0–8). VEN was administered with azacitidine (AZA) in 59%, with decitabine (DEC) in 29% and with low-dose cytarabine (LDAC) in 12% of patients, respectively. The CR/CRi and PR rates were 12.4% and 10.4%. The CR/CRi and overall response (OR, CR/CRi+PR) were higher in patients receiving VEN+AZA (17.9% and 32.1%) than in those receiving DEC + VEN (6.7% and 13.3%) or LDAC + VEN (0%). The presence of mutated NPM1 and monoallelic CEBPA were the only two variables associated with increased CR/CRi with VEN in AML-RR. Median OS was 104 days (95% CI: 56–151). Treatment response and ECOG 0 were the only variables that influenced OS in a multivariate model adjusted for age and sex. Patients who received subsequent salvage therapy had superior OS (98 vs 5 days, p=0.004).Twenty-eight percent of patients required discontinuation of VEN due to toxicity; 61% of patients required hospital admission, mainly due to infections (45%). Conclusions: The observed responses and OS were far from the results observed in first-line and inferior to most retrospective series in RR-AML due to patients’ poor prognostic features but were very similar to another series with similar characteristics, being an option for RR-AML.