PI-074 - MODEL-BASED DRUG DEVELOPMENT CONSIDERATIONS: ZANIDATAMAB IN PATIENTS WITH HER2 EXPRESSING BILIARY TRACT CANCER (BTC)

L. Wu¹, S. Trueman¹, S. Girgis¹, D. Jaworowicz², J. Passarell², R. Iannone¹, J. Ma¹, H. Zhou¹; ¹Jazz Pharmaceuticals, Palo Alto, CA, USA, ²Simulations-Plus, Buffalo, NY, USA.

Background: Zanidatamab is a HER2-targeted bispecific antibody that binds to 2 non-overlapping HER2 domains. The population pharmacokinetic (PPK) and exposure-response (ER) analyses for confirmed objective response (cOR), progression-free survival (PFS) and selected treatment-emergent adverse events (TEAEs) from 2 monotherapy studies were used to support the BLA for zanidatamab in BTC.

Methods: The PPK and ER safety included phase 2 HERIZON-BTC-01 (NCT04466891; n = 87, dose 20 mg/kg Q2W) and phase 1 ZW25 101 (NCT02892123; n = 192, dose ranged from 5 to 30 mg/kg including QW, Q2W, and Q3W), while ER for cOR and PFS included only HERIZON-BTC-01. The PPK model was developed using nonlinear mixed-effects modeling (NONMEM). The exposure metrics for ER analyses were derived based on the PPK model.

Results: A total of 6562 serum PK from 279 patients were included. BTC was the most prevalent cancer type (39.1%), with breast cancer, GEA, and CRC comprising 19.0%, 17.9%, and 10.4% of the analysis population, respectively. The final PPK model was a 2-compartment model with parallel linear and nonlinear CL. Covariate effects identified were body weight, albumin, number of lesions, and cancer type on linear CL; body weight and cancer type on Vc; and cancer type on Vp. None of the statistically significant covariate effects result in clinically relevant differences in steady-state exposures, except for patients with GEA (AUC0-t,ss geometric mean ratio of 0.722 relative to patients with BTC). The final ER cOR model was a logistic regression with a power function of Cycle 1 Cmin with the majority of participants having an exposure range that was on the plateau region of the ER curve. The final ER PFS model was a Cox proportional hazards model that included the effect of daily Cavg and IHC status (3+ versus 2+/1+/0). No statistically significant ER safety relationship was found for Grade ≥ 3 diarrhea, Grade ≥3 TEAEs or IRRs. Both exposure-efficacy and exposure-safety analyses support the 20 mg/kg Q2W dose selection.

Conclusion: The model-based analyses support the recommended zanidatamab 20 mg/kg Q2W as a dose that provides promising efficacy with manageable safety in BTC.