(T-037) Exposure-Response Efficacy and Safety Analysis of Repotrectinib to Support the Dose Recommendation for Patients with ROS1-positive NSCLC

Lora Hamuro, na – Senior Director, BMS; Zheyi Hu, NA – Director, BMS; Anna Kondic, NA – Executive Director, BMS; Ming Lu, NA – NA, NA; Amit Roy, na – Head consulting, PUMAS; Jun Shen, na – Director, BMS; Li Zhu, NA – executive director, BMS

Objectives: Repotrectinib is a tyrosine kinase inhibitor approved in the US for the treatment of adult patients with locally advanced or metastatic ROS1-positive NSCLC and currently being developed for the treatment of adult and pediatric patients with NTRK-positive solid tumors. Exposure - response (E-R) models were developed with selected key efficacy endpoints (OR: objective response rate and PFS: progression-free survival) and clinical safety endpoints (Gr2+ dizziness: ≥ Grade 2 dizziness, DRDI: dose reduction or interruption due to AEs) to support the benefit-risk assessment of repotrectinib and dose justification for the treatment of patients with ROS1-positive NSCLC.

Methods: Data from TRIDENT-1 (NCT03093116) were used to evaluate E-R relationships. A logistic regression model was used to understand E R of objective response with respect to time-averaged repotrectinib concentration over the first 56 days of dosing (Cavg56). Cox Proportional-Hazards (CPH) models were used to characterize the relationship of PFS as well as safety endpoints (Gr2+ dizziness, DRDI) with respect to time-varying cumulative repotrectinib concentration (Cavgc) averaged over every 12-hours . The logistic regression model and CPH model were used to predict and compare the efficacy and safety for dosing regimens 160 mg QD/BID (160 mg QD for 14 days, if tolerated, followed by 160 mg BID) and 160 mg QD under different food status.

Results: The probability of OR or PFS for ROS1-positive NSCLC increased with exposure. As compared with 160 mg QD, the recommended dose regimen of 160 mg QD/BID was predicted to show improved efficacy (eg, ORR 85.0% vs 76.1% and 46.6% vs 33.1% for TKI-naive and TKI-pretreated respectively) but with only a small increase in AEs (eg, Gr2+ dizziness 15% vs 13%). The model-predicted probabilities of all evaluated efficacy endpoints and AEs are similar among the different food conditions and supports repotrectinib dosing without regard to food status. This work emphasizes the importance of selecting an appropriate exposure for the E-R analyses when dose or dose frequencies change throughout treatment.

Citations: NA