Systematic literature review and indirect treatment comparisons (ITC) of glasdegib (GLAS) plus low dose ara-c (LDAC) versus a hypomethylating agent (HMA) for previously untreated acute myeloid leukemia (AML) patients ineligible for intensive chemotherapy (NIC).

Abstract

e18526 Background: In a randomized phase 2 study, GLAS, an oral, smoothened inhibitor, combined with LDAC, showed significantly better overall survival (OS) vs LDAC alone in previously untreated AML NIC patients. HMAs azacitidine (AZA) and decitabine (DEC) are considered current standard of care in this population. An ITC was conducted to compare OS for GLAS+LDAC vs. AZA and DEC, respectively. Methods: Embase, MEDLINE, Cochrane database, and conference abstracts (ASCO, ESMO and ASH) were systematically searched through 12/2016 for relevant RCTs of GLAS, AZA and DEC in AML patients ineligible for IC. Classical frequentist ITC using the Bucher method was used to indirectly compare OS hazards ratios with 95% confidence intervals (CI) using LDAC as the common comparator. Results: Four studies met inclusion criteria: AZA Fenaux 2010, N (treatment/ comparator) = 14/20, AZA Dombret 2015, N = 241/158, DEC: Kantarjian 2012, N = 242/243, and GLAS+LDAC: Cortes 2016, N = 88/44. Upon review of baseline characteristics, Fenaux 2010 was excluded based on major population differences (% bone marrow blasts). Three studies contributed data to ITC based on comparable populations: age, cytogenic risk and OS of comparator were similar among studies: age 75/73/76 years old, poor cytogenic risk 34%/37%/39%, OS of the comparator 6.4/5.0/4.3 months in AZA/DEC/GLAS+LDAC, respectively. In the ITC, GLAS+LDAC showed significantly better OS HR vs. AZA and DEC (0.51 and 0.56 respectively) (Table). Conclusions: Using ITC, treatment with GLAS+LDAC yielded significantly better OS HR than AZA and DEC in previously untreated AML patients ineligible for treatment with IC. Limitations include mixed IC & NIC population for the AZA trial, and mixed comparator arm of both LDAC and BSC for the DEC trial. Analyses using patient-level data matching baseline characteristics across studies may enable more robust ITC. [Table: see text]