Exposure-response relationship of regorafenib efficacy in patients with hepatocellular carcinoma

To explore the relationship between regorafenib exposure and efficacy in patients with hepatocellular carcinoma (HCC) who had disease progression during sorafenib treatment (RESORCE).

Exposure-response (ER) analyses for regorafenib were performed using data from a phase 3, randomized, placebo-controlled trial (RESORCE). Patients received 160mg regorafenib or placebo once daily (3weeks on/1week off in a 4-week cycle) with best supportive care until disease progression, death, or unacceptable toxicity. Kaplan-Meier analyses for overall survival (OS) and time-to-progression (TTP) were performed in which regorafenib-treated patients were grouped into four categories according to their estimated average exposure over 4weeks in cycle 1. While this analysis primarily focused on efficacy, a potential correlation between exposure and treatment-emergent adverse events (TEAEs) was also evaluated. If any differences were observed between Kaplan-Meier plots, the ER analysis continued with a multivariate Cox regression analysis to evaluate the correlation between exposure quartile categories and the efficacy and safety parameters while taking into consideration the effect of the predefined clinically relevant demographic and baseline covariates. The functional form of the ER relationship within the regorafenib treatment group was subsequently evaluated.

Based on visual assessment of the Kaplan-Meier plots, no meaningful relationship between the exposure categories and TEAEs were observed, although median OS and TTP tended to be longer in the higher exposure categories. Further ER analyses, which considered the effects of predefined covariates and the different shapes of the ER relationship, focused on efficacy. The baseline risk factors Eastern Cooperative Oncology Group (ECOG) performance status ≥1, alpha-fetoprotein levels ≥400ng/ml, and aspartate transaminase or alanine transaminase levels >3×upper limit of normal were significantly associated with OS (P<0.01) and age was associated with TTP. A statistically significant difference was found for OS and TTP between patients receiving regorafenib compared with those receiving placebo in the multivariate ER analysis (P<0.01) in favor of regorafenib. However, within the group of regorafenib-treated patients, the effect of regorafenib exposure on efficacy, either by estimating four effect sizes for each quartile, or by including a continuous linear or nonlinear relationship between individual exposure and efficacy, was not significant (P>0.01) and relatively flat. This suggests that increasing regorafenib exposure would not result in a meaningful increase in OS or TTP.

After considering the baseline risk factors: ECOG performance status, alpha-fetoprotein levels, and hepatic function for OS and age for TTP, the ER analysis in regorafenib-treated patients showed similar efficacy over the entire predicted exposure range in RESORCE. This supports the selected regorafenib dose of 160mg once daily (3weeks on/1week off in a 4-week cycle) in patients with intermediate or advanced HCC who have experienced disease progression on sorafenib.