Matching-adjusted indirect comparison (MAIC) of olutasidenib (OLU) and ivosidenib (IVO) in isocitrate dehydrogenase 1 (IDH1)-mutated relapsed/refractory (R/R) acute myeloid leukemia (AML).

Abstract

6546 Background: OLU and IVO are allosteric type II IDH1 inhibitors approved by the FDA and recommended by NCCN for IDH1 mut R/R AML patients based on single-arm trials. In the absence of a head-to-head trial, a MAIC was performed to estimate relative treatment effects of OLU vs. IVO in IDH1 mut R/R AML. Methods: Analyses used registrational data for OLU (Study 2102-HEM-101; N=147; individual-level data) and IVO (AG120-C-001; N=174; study-level data). A logistic propensity score model was used to estimate weights based on the first moment for Study 2102-HEM-101 patients to match AG120-C-001, including the following characteristics identified from a literature review, validated by clinical experts: number of prior systemic therapies, age, prior stem cell transplant, AML type, relapse type, cytogenetic risk, ECOG PS, and IDH1 mutation. Complete remission (CR) and CR + CR with partial hematological recovery (CRh) were summarized as odds ratios (ORs) and 95% confidence intervals (CIs). Duration of CR (DoCR), duration of CR+CRh, and OS were summarized in terms of difference in medians and 95% CIs. OS was also summarized in terms of hazard ratios (HRs) and restricted mean survival time (RMST). A simulated treatment comparison (STC) was performed as a sensitivity analysis. Results: Table 1 summarizes MAIC-adjusted estimates. Naïve and adjusted rates of CR and CR+CRh for OLU vs. IVO were comparable, but point estimates favored OLU for CR. Differences in median DoCR were not statistically significant but favored OLU over IVO. OLU had a significantly longer duration of CR+CRh than IVO. For OS, the naïve comparison suggested OLU was better than IVO (HR=0.72; 95% CI 0.55, 0.92), whereas the MAIC was uncertain but favored OLU. STC results were consistent with the MAIC. Conclusions: Naïve and adjusted rates of response for OLU vs. IVO were comparable (adjusted point estimate favored OLU for CR and IVO for CR+CRh), while a longer duration of CR+CRh was observed with OLU. Adjusted OS was similar between the two groups, although the HR favored OLU, and could not be estimated by response category given lack of patient characteristics and reduction in effective sample size (ESS). Results rely on the assumption of no unmeasured confounders which reflects a limitation of the methodology. MAIC results. Outcome OLU – adjusted (95% CI) IVO – observed (95% CI) MAIC OLU vs. IVO (95% CI) N ESS CR 27% 25% OR=1.12 (0.61, 2.08) 147 73.03 CR + CRh 29% 33% OR=0.83 (0.46, 1.50) 147 73.03 DoCR, median mos 21.3 (12.0, NE) 10.1 (6.5, 22.2) Diff=11.18 (-4.30, 22.72) 47 17.76 Duration of CR+CRh, median mos 17.5 (12.0, 29.1) 8.2 (5.6, 12.0) Diff=9.84 (3.24, 22.28) 51 21.06 OS, median mos 9.7 (5.6, 16.4) 9.0 (7.4, 10.2) HR=0.75 (0.53, 1.07) 147 73.03 OS, RMST mos 15.6 (12.1, 19.2) 12.2 (10.6, 13.9) Diff=3.39 (-0.51, 7.29) 147 73.03 N, sample size; NE, not estimable.