OBJECTIVES: Vismodegib is a Hedgehog (Hh) pathway inhibitor indicated for the treatment of adults with advanced basal cell carcinoma (aBCC). Vismodegib is licensed for two distinct subgroups: symptomatic, metastatic BCC (mBCC) and locally advanced BCC (laBCC) deemed inappropriate for surgery or radiotherapy. Registration studies for vismodegib were both uncontrolled and single arm (ERIVANCE [2008-004945-27], STEVIE [NCT01367665]), which poses a significant challenge when populating the clinical aspect of the comparator arm in an economic model. The objective of this study was to estimate the comparative effectiveness of vismodegib against best supportive care (BSC) using a “Landmark” approach, and use the results to inform an economic evaluation of vismodegib in aBCC.

METHODS: Non-responders at 6 months in STEVIE were used as a proxy group for patients receiving BSC. A 6-month landmark was selected as it exceeds the mean and median time to first confirmed response. A Cox regression model was used to estimate progression-free survival (PFS) and overall survival (OS) hazard ratios (HRs) of non-responders versus responders. Hazard ratios were estimated using both a common and separate treatment effect across the laBCC and mBCC populations with and without adjustment for age, and ECOG status.

RESULTS: Mean [95% CIs] HRs estimated using a 6-month landmark and the common treatment effect amounted to 2.16 [1.27, 3.68] and 1.31 [0.99, 1.75] for OS and PFS, respectively. The separate treatment effect model yielded similar and clinically plausible HRs in the locally advanced population. In the metastatic population, HRs were not statistically significant because of the low number of patients and events.

CONCLUSIONS: The landmark method removes bias that would otherwise occur in a comparison of ever-responders and never-responders over the entire observation period. Despite limitations, the landmark approach can be considered robust methodology when deriving comparative effectiveness from uncontrolled trials.