Comparative Efficacy of Targeted Therapies in Old Unfit Intermediate to High-Risk Patients with Acute Myeloid Leukemia: A Meta-Analysis of Comparative Studies

Introduction: Acute myeloid leukemia (AML) commonly affects the elderly with a median age of 67 years at diagnosis. Elderly AML is associated with high risk factors and adverse cytogenetics contributing to its worse outcomes. Moreover, older patients with AML respond poorly to intensive induction chemotherapy due pre-existing comorbidities. Novel therapies that include Bcl-2 inhibitors (Bcl-2i) and hypomethylating agents (HMA) have been used to treat AML in this patient population with varying degrees of response and success. Often, combination therapy with Bcl-2i and HMA has been used as frontline therapy for older patients or younger unfit patients who are not candidates for intensive chemotherapy. Currently, there are no randomized studies comparing various HMA, Bcl-2i, IDH-targeting agents, and their combinations. We performed a set of meta-analyses to evaluate relative efficacy of these various regimens in the treatment of AML in an old unfit intermediate to high-risk patient population. Methods: A review of the medical literature was conducted using online databases. Inclusion criteria consisted of English language; newly diagnosed or relapsed/refractory (R/R) AML; AML in elderly, unfit, intermediate to high-risk patients; comparative studies involving Azacitidine (A), Decitabine (D), Venetoclax (V), Enasidenib (E), and intensive chemotherapy (IC); and comparative studies reporting death rates over 2 year-period. Relative risk (RR) of death was calculated for each regimen. For comparative efficacy of different therapeutic regimens, a frequentists network meta-analysis was conducted using netmeta package and random effects model. For molecular alteration % response across various studies, a proportional meta-analysis was conducted. The pooled proportion with 95% CI was given both for the fixed effects model and the random effects model. Results: Five studies comprising 1,109 participants were included. The RR of death was lowest in A+V<D+V<E+A<A<IC in increasing order. This pattern persisted when the data was analyzed by RR of death at 1, 1.5, and 2 years. While A+V RR of death was the lowest at 1.5 years compared to the rest of the treatment regimens, theirs did not seem to vary much at the different time points. A+V and D+V had a significantly superior RR of death compared to IC, whereas A+V demonstrated also superior RR of death over A. The network meta-analysis ranked A+V as the most favorable regimen for newly diagnosed or R/R AML in an unfit, intermediate to high-risk patients followed by D+V, E+A, A, and IC in a decreasing order. We then analyzed the % CR/CRi rates of A+V across different AML molecular alternations. Those with CR/CRi rates <30%: KRAS, CBL, U2AF1. Alterations with CR/CRi rates 30-49%: JAK2, ETV6, GATA2, ZRSR2, ASXL1, CEBPA, SRSF2, STAG2, WT1. Those with CR/CRi rates of 50-70%: BCOR, BCORL, NF1, TP53, FLT3ITD, TET2, NRAS, DNMT3A, FLT3TKD, PTPN11, RUNX1, SF3B1, IDH2, EZH2. Lastly alterations with CR/CRi rates > 70%: IDH1, NPM1. Conclusions: This network meta-analysis is the first to compare and rank AML targeted therapies in patients with newly diagnosed or R/R AML who are intermediate to high-risk and unfit. It indicates that A+V has better RR of death than A or IC and was ranked as the preferred combination. It also demonstrates that A+V has wide range of activity with demonstrable CR/CRi across various AML molecular alterations.

No relevant conflicts of interest to declare.