Efficacy of regorafenib (REG) in patients with hepatocellular carcinoma (HCC) in the phase III RESORCE trial according to alpha-fetoprotein (AFP) and c-Met levels as predictors of poor prognosis.

Abstract

4078 Background: REG is a multikinase inhibitor which improved overall survival (OS; HR 0.63, 95% CI 0.50, 0.79; P<0.0001) and time to progression (TTP; HR 0.44, 95% CI 0.36, 0.55; P<0.0001) compared with placebo in patients with HCC who progressed during prior sorafenib treatment in the RESORCE trial. This exploratory analysis evaluated the impact of baseline AFP and c-Met on REG treatment benefit (OS and TTP) in the RESORCE trial. Methods: Circulating AFP and c-Met protein (shed ectodomain) levels were quantified by a Luminex assay (Myriad RBM) in plasma samples collected at baseline from patients enrolled in the RESORCE trial. Valid biomarker data were available from 497 (AFP) and 499 (c-Met) out of 573 patients. Patients were subgrouped according to the median protein concentration (high vs low), and the treatment effect HR and its 95% CI were evaluated using a Cox proportional hazards model. The predictive effect was modeledas a protein–treatment interaction effect and subjected to Akaike information criterion (AIC)-based selection to assess its association with OS and TTP. Results: Baseline characteristics of patients were balanced across protein subgroups. While increased levels of both AFP (HR 1.09, 95% CI 1.07, 1.12; P<0.001) and c-Met (HR 1.32, CI 95% 1.06, 1.63; P=0.011) were associated with a worse prognosis for OS, increased AFP levels were also associated with poor prognosis for TTP (HR 1.05, 95% CI 1.03, 1.07; P<0.001). REG treatment benefit for both OS and TTP was independent of AFP and c-Met protein expression (Table). The protein–treatment interaction effect was not statistically significant. Conclusions: The treatment benefit of REG in patients with HCC was independent of AFP and c-MET protein expression at baseline. So far, no single protein has been associated with REG clinical benefit. Clinical trial information: NCT01774344. [Table: see text]