Background: Glasdegib (GLAS) is an oral smoothened inhibitor that showed significantly better overall survival (OS) when combined with Low Dose ARA-C (LDAC) versus LDAC alone in a randomized Phase 2 study of previously untreated acute myeloid leukemia (AML) patients ineligible for intensive chemotherapy (NIC).In this trial, patients receivingGLAS+LDAC versus LDAC alone showed imbalance in baseline characteristics, which could impact the results of an indirect treatment comparison (ITC). Hypomethylating agents (HMAs) azacitidine (AZA), and decitabine (DEC) are considered current standard of care in this population. There are no head-to-head studies comparing these treatments to GLAS+LDAC.

Aims: To adjust efficacy results for baseline covariates in this indirect treatment comparison analysis, and enable a more robust comparison of GLAS+LDAC to AZA and DEC.

Methods: Individual patient data from the GLAS+LDAC treatment group were adjusted for baseline characteristics to closely resemble the LDAC group in the Phase 2 study: mean age, gender, de novo disease, ≥50% bone marrow blasts, cytogenetic risk, Eastern Cooperative Oncology Group (ECOG) performance status, and mean hemoglobin (HgB) were selected for adjustment. Although none of the baseline variables were identified as statistically significant predictors of OS, the male gender, de novo disease, and ≥ 50% bone marrow blasts were associated with a negative impact on OS. Age, ECOG status, and mean HgB did not demonstrate a meaningful impact on OS. Cytogenetic risk was a stratification variable in the GLAS randomized controlled trial and was not expected to have an impact on OS in this analysis. Classical frequentist ITC implementing the Bucher method was then used to indirectly compare covariate adjusted OS hazards ratios (HRs) with 95% confidence intervals (CI) using LDAC as the common comparator.

Results: The intent-to-treat (ITT) OS Cox model for GLAS+LDAC vs. LDAC demonstrated OS HR=0.463 (0.299-0.717). The covariate adjustment lead to OS HR=0.443 (0.275-0.712). ITC of GLAS+LDAC vs. AZA yielded OS HRs of 0.514 (0.310-0.852) and 0.492 (0.287-0.843) for ITT and covariate adjusted OS, respectively. ITC of GLAS+LDAC vs. DEC yielded OS HRs of 0.564 (0.351-0.908) and 0.540 (0.324-0.900) (Table).

Conclusions: The results of the covariate adjusted ITC confirmed the robustness of the ITT analysis and demonstrated statistically significant improvements in OS for GLAS+LDAC as compared to AZA and DEC.

CHART: Covariate Adjusted ITC: GLAS+LDAC vs. AZA and GLAS+LDAC vs. DEC


Arondekar: Pfizer: Employment, Equity Ownership. Chan: Pfizer Inc: Employment. Yun: Pfizer: Employment.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution