OBJECTIVES: The National Institute for Health and Care Excellence recommends two methods for adjusting between-trial population imbalances in indirect treatment comparisons (ITC): Matching-Adjusted Indirect Comparison (MAIC) and Simulated Treatment Comparison (STC). While neither method is recommended more, this case study compares both approaches plus standard (unadjusted) ITC among patients ineligible to receive intensive chemotherapy.

METHODS: Using standard ITC, MAIC, and STC, results of the Phase II glasdegib with low-dose ARA-C (GLAS+LDAC) trial (n=116) were indirectly compared to published Phase III azacitidine (AZA) trial data, with LDAC alone as the common comparator. In MAIC, patient-level data for GLAS+LDAC were weighted to match mean baseline characteristics reported for AZA. In STC, GLAS+LDAC data generated a regression model with baseline characteristics as covariates, which was used to simulate outcomes for AZA trial participants. Overall survival (OS) hazard ratios (HR) with 95% confidence intervals (CIs) were estimated.

RESULTS: Standard ITC demonstrated GLAS+LDAC superiority over AZA (HR=0.57; 95%CI: 0.35-0.91). Using MAIC, propensity score weighting reduced effective sample size to 32 (72% loss). MAIC estimated improved OS in favor of GLAS+LDAC, but did not reach statistical significance (HR=0.87; 95%CI: 0.48-1.58). In STC, adjusting for key population covariates found a similar yet stronger, more precise survival effect (HR=0.47; 95%CI: 0.26-0.85) without reducing sample size.

CONCLUSIONS: In each ITC, GLAS+LDAC is associated with improved OS. Preserving sample size is important for subpopulations to minimize uncertainty around point estimates, improve model robustness, and derive more applicable results. While standard ITC and STC preserve the sample, only STC enables population-specific interpretations. In MAIC, significant results and interpretations are severely limited by sample size loss. Although the literature reflects increasing MAIC use despite its limitations regarding sample size reduction, choosing an ITC method should be guided by characteristics of data available to ensure robust analyses and appropriate interpretation of the data.