Analysis of survival and objective response (OR) in patients with hepatocellular carcinoma in a phase III study of lenvatinib (REFLECT).

186 Background: In REFLECT, Lenvatinib (LEN) demonstrated treatment effect on overall survival (OS) by statistical confirmation of noninferiority to sorafenib (SOR). OR rates for LEN versus SOR were: 24% versus 9% by investigator review and 41% versus 12% by independent review (Kudo et al 2018). Since the relationship between OR and OS in phase III HCC studies is unclear, we explored the relationship between OR and OS in REFLECT. Methods: OR assessed by investigators per mRECIST were used to analyze the association between OR and OS of pts treated with LEN or SOR. The median OS of responders (CR or PR) was compared to that of nonresponders (SD, PD, or UNK/NE) irrespective of treatment. Landmark analyses were performed by OR status at several fixed time points as sensitivity analyses, and the effect on OS was evaluated by Cox regression with OR as a time-dependent covariate, with other prognostic factors. Results: Median OS was 22.4 months for responders and 11.4 months for nonresponders. Hazard ratios (HR) of landmark analyses at 2, 4, and 6 months were 0.75 (95% CI, 0.57–0.98), 0.72 (95% CI, 0.56–0.92), and 0.73 (95% CI, 0.57–0.93). Independent predictors of OS based on unstratified Cox regression are in the table. Conclusions: In REFLECT, OR was an independent predictor of OS in pts with HCC regardless of treatment. The results indicate this correlation is worth further investigation. Clinical trial information: NCT01761266. [Table: see text]