A Markov analysis of azacitidine and venetoclax vs induction chemotherapy for medically fit patients with AML

Abstract

Abstract Although induction chemotherapy (IC) is the standard of care in medically fit patients with newly diagnosed acute myeloid leukemia (AML), limited retrospective data indicate that patients at adverse-risk may benefit from azacytidine and venetoclax (aza-ven). Our goal was to perform a Markov decision analysis to determine whether IC or aza-ven is the optimal induction regimen in this population. Using the TreeAge software, Markov models were created for adverse-risk and intermediate-risk cohorts. A systematic review of the literature informed the transition probabilities and utilities included in the analyses. Our primary outcome was quality-adjusted life years (QALYs) gained over 5 years after diagnosis. Overall, patients at adverse risk treated with IC gained 1.4 QALYs, compared with 2.0 QALYs in patients treated with aza-ven. Patients at adverse risk treated with IC and allogeneic stem cell transplantation (allo-SCT), IC, aza-ven and allo-SCT, or aza-ven gained 2.1, 1.5, 3.0, and 1.9 QALYs, respectively. Meanwhile, patients at intermediate risk treated with IC gained 2.0 QALY, compared with 1.7 QALY in patients treated with aza-ven. Patients at intermediate risk treated with IC and allo-SCT, IC, aza-ven and allo-SCT, and aza-ven gained 2.7, 2.3, 2.6, and 1.8 QALYs, respectively. We have demonstrated that medically fit patients with newly diagnosed adverse-risk AML may benefit from treatment with aza-ven over those treated with IC, whereas IC remains the preferred approach for patients at intermediate risk. Our work challenges the use of the European LeukemiaNet risk classification for patients treated with aza-ven and highlights the need for prospective investigation into aza-ven as induction therapy for medically fit patients.