(W-112) Use of Model-Informed Drug Development (MIDD) to Support Dose Selection of Nivolumab in Combination with Ipilimumab in Adolescent Patients with MSI-H/dMMR Metastatic Colorectal Cancer

Jian Shi, PhD – Senior Research Investigator, BMS; Jian Zhou, PhD – Associate Director, BMS; Yue Zhao, PhD – Director, BMS; Mayu Osawa, PhD – Associate Director, BMS; Paul Statkevich, PhD – Senior Director, BMS; Anna Kondic, PhD – Executive Director, BMS; Li Zhu, PhD – Executive Director, BMS; Ada Zhuang, PhD – Director, BMS

Objectives: Nivolumab (nivo) in combination with ipilimumab (ipi) is approved in the US for adults and adolescents (12 to < 18 years) with melanoma and 3L+ MSI-H/dMMR colorectal cancer (CRC). A MIDD approach was used to support the proposed dosing regimen in adolescents (weighing ≥ 40 kg: nivo 240 mg + ipi 1 mg/kg Q3W x 4 doses, then nivo 240 mg Q2W or 480 mg Q4W; weighing < 40 kg: nivo 3 mg/kg + ipi 1 mg/kg Q3W x 4 doses, then nivo 3 mg/kg Q2W) with 1L MSI-H/dMMR CRC by comparing the simulated exposures between adolescents and adults based on popPK models.

Methods: The developed nivo/ipi pediatric popPK models were updated to include PK data in adults with 3L+ CRC from Study CheckMate142. The popPK models were validated by goodness-of-fit and pcVPC plots, then used for stochastic simulations of nivo/ipi PK in adults (from CheckMate142 with covariates of age, gender, race, body weight, lean body mass, PS, eGFR) and adolescents (randomly sampled from the NHANES database with covariates of gender and body weight; other covariates were estimated by equations or assigned based on popPK dataset) with CRC. The simulation results were grouped into a series of body weight bands. The simulated geometric mean exposure measures at 1st, 4th, and steady-state (Cavg1, Cmin1, Cmax1, Cavg4, Cmin4, Cmax4, Cavgss [nivo only], Cminss [nivo only], and Cmaxss [nivo only]) dosing intervals in adolescents for each weight band were compared to the adult reference range of geometric means of exposure for the spectrum of weights.

Results: The PK parameter estimates of nivo and ipi based on the popPK models were consistent with the prior analysis. The goodness-of-fit plots showed good agreement between the observed and predicted PK data. The pcVPC plots showed that the models adequately characterized the data. The simulation results showed nivo/ipi exposures in adolescents were generally within the adult reference ranges. Although nivo exposures in some lower body weight bands appeared higher in adolescents than adult reference ranges (up to 20.3-38.7% higher), and ipi exposures in several higher body weight bands were higher in adolescents than adult reference ranges (up to 21.7-61.2% higher), this is not clinically significant given: 1) predicted adolescent nivo exposures were below those achieved in adults receiving nivo 10 mg/kg Q2W, which was safe and tolerable; 2) predicted adolescent ipi exposures were below those achieved in adults with adjuvantly-treated melanoma receiving the approved 10 mg/kg Q3W dose; 3) previously demonstrated flat/shallow E-R efficacy or safety relationships for both drugs across tumor types.


Conclusions: The exposures of the proposed dosing regimen for both nivo and ipi in adolescents are expected to result in comparable benefits and risks to those in adults. MIDD is a powerful tool to support dose recommendations in adolescent populations with limited clinical data.

Citations: NA