OBJECTIVES: Based on single-arm trial-data (BOLT; Migden et al., 2015), sonidegib was approved recently in the US and EU to treat locally advanced basal cell carcinomas (laBCCs) ineligible for curative surgery or radiotherapy. Vismodegib, the other approved targeted therapy for advanced BCC, was also assessed in a single- arm trial (ERIVANCE; Sekulic et al., 2012). Our objective was to examine the comparative effectiveness of the two drugs using a matching-adjusted indirect comparison (MAIC; Signorovitch et al., 2010, 2012) versus an unadjusted indirect comparison.
METHODS: After comparing study designs, patient baseline characteristics, and outcome definitions, and conducting a targeted literature review and consulting with clinical advisors to identify potential prognostic factors, we used an MAIC to adjust for differences in key patient baseline characteristics. Due to the small relevant BOLT sample size (n= 66), the number of matching variables was restricted to two. BOLT patient-level data (18 months of follow-up; Novartis data on file, 2015) were weighted such that the two key baseline patient characteristics matched published values for laBCC patients in ERIVANCE (12-month update, i.e., 21 months of follow-up; Sekulic et al., 2015). Efficacy results for sonidegib were generated using matched patient-level data and compared to published results from ERIVANCE.
RESULTS: Matching variables were baseline prevalence of prior radiotherapy (7.6% [BOLT] vs 20.6% [ERIVANCE]) and prior surgery (72.7% [BOLT] vs 88.9% [ERIVANCE]). The matching procedure was effective (BOLT variables post-matching: 20.6% and 89.0%, respectively). After re-weighting, sonidegib objective response rate (ORR) and median progression-free survival (PFS) were effectively unchanged (sonidegib pre- vs post-matched ORR and PFS were 56.1% vs 56.7% and 22.1 vs 22.1 months). Vismodegib ORR and PFS were 47.6% and 9.5 months.
CONCLUSIONS: The comparative effectiveness of sonidegib vs vismodegib remains unchanged after adjusting BOLT patient-level data to match published ERIVANCE values for baseline prevalence of prior surgery and radiotherapy.
