OBJECTIVES

: To assess cost-effectiveness (CE) of regorafenib in comparison with best-supportive care (BSC) for previously treated patients with metastatic colorectal cancer (mCRC) in the Czech Republic (CZ) based on national registry data.

METHODS

: Regorafenib efficacy data were mandatorily collected in registry between 07/2015 and 07/2017 during 3-year temporary (innovative) reimbursement (n=429). BSC efficacy data were based on randomized clinical trial (RCT; CORRECT) of regorafenib/BSC with unpublished additional 2.5-year follow-up. Individual data on progression-free survival (PFS)/overall survival (OS) were adjusted/balanced for age, gender, ECOG performance status, number of treatment lines, mCRC localization, KRAS mutation, time from metastases using propensity score weighting (PSW) with overlap weights estimated by logistic regression.

The weighted Kaplan-Meier PFS/OS curves were put into the health economic Markov model with three health states: progression-free/ progression/ death. The costs were based on list prices, reimbursement tariffs and expert opinion as of 09/2017. Costs and outcomes were discounted by 3%. Probabilistic sensitivity analysis (PSA) with 3,000 iterations using an implicit CZ willingness-to-pay threshold (WTP) equal to €47,000/QALY was performed.

RESULTS

: The registry data showed improved survival compared to RCT despite PSW: hazard ratio (HR) for OS was newly 0.53 (95%CI: 0.34-0.83) and for PFS 0.32 (95%CI: 0.21-0.49) (RCT HR for OS: 0.79 (95%CI: 0.67-0.92), for PFS 0.46 (95%CI: 0.39-0.55)).

Over a life-time horizon, regorafenib had an incremental cost-effectiveness ratio of €43,122/QALY. Regorafenib brought 0.221 incremental QALY (0.572 vs. 0.351) and incremental costs of €9542 (€9904 vs. €362). PSA showed that probability of regorafenib being cost-effective is 81% at WTP.

CONCLUSIONS

: Regorafenib is a cost-effective therapy for patients with mCRC based on registry/RCT data in the CZ. To our knowledge, this is the first CE analysis of regorafenib using extensive registry data with PSW method and the first CE analysis for regorafenib in CZ. Nevertheless, PSW balances characteristics, but does not substitute randomization.