Repotrectinib Exhibits Potent Antitumor Activity in Treatment-Naïve and Solvent-Front-Mutant ROS1-Rearranged Non-Small Cell Lung Cancer.

PURPOSE: Although first-line crizotinib treatment leads to clinical benefit in ROS1+ lung cancer, high prevalence of crizotinib-resistant ROS1-G2032R (ROS1G2032R) mutation and progression in the central nervous system (CNS) represents a therapeutic challenge. Here, we investigated the antitumor activity of repotrectinib, a novel next-generation ROS1/TRK/ALK-tyrosine kinase inhibitor (TKI) in ROS1+ patient-derived preclinical models. EXPERIMENTAL
DESIGN: Antitumor activity of repotrectinib was evaluated in ROS1+ patient-derived preclinical models including treatment-naïve and ROS1G2032R models and was further demonstrated in patients enrolled in an on-going phase I/II clinical trial (NCT03093116). Intracranial antitumor activity of repotrectinib was evaluated in a brain-metastasis mouse model.
RESULTS: Repotrectinib potently inhibited in vitro and in vivo tumor growth and ROS1 downstream signal in treatment-naïve YU1078 compared with clinically available crizotinib, ceritinib, and entrectinib. Despite comparable tumor regression between repotrectinib and lorlatinib in YU1078-derived xenograft model, repotrectinib markedly delayed the onset of tumor recurrence following drug withdrawal. Moreover, repotrectinib induced profound antitumor activity in the CNS with efficient blood-brain barrier penetrating properties. Notably, repotrectinib showed selective and potent in vitro and in vivo activity against ROS1G2032R. These findings were supported by systemic and intracranial activity of repotrectinib observed in patients enrolled in the on-going clinical trial.
CONCLUSIONS: Repotrectinib is a novel next-generation ROS1-TKI with improved potency and selectivity against treatment-naïve and ROS1G2032R with efficient CNS penetration. Our findings suggest that repotrectinib can be effective both as first-line and after progression to prior ROS1-TKI. ©2020 American Association for Cancer Research.