041: Population Modeling Framework With Adaptive Dosing for Optimizing Clinical Utility of Compounds in Development for Treatment of Solid Tumors

Statement of

Purpose: The aim was to develop a multi-model-informed drug development (MIDD) framework for informing and optimizing the dosing regimens of oncology compounds in early to mid-phase development that are intended to treat solid tumors, through incorporation of the population models of pharmacokinetics, tumor dynamics and longitudinal safety profile (e.g. myelosuppression) with consideration of expected dose modifications, ultimately expressed in simplified scores of clinical utility.

Description of Methods & Materials:

Methods: Population pharmacokinetic (popPK) models and longitudinal population exposure-response (ER) models for tumor dynamics and myelosuppression were developed as fit-for-purpose and refined as data arose throughout early-stage development of novel compounds. Predictions of objective response rate (ORR) and incidence of hematologic adverse events by severity were derived from longitudinal models across various starting doses. Additionally, other logistic regression ER models for efficacy and/or safety endpoints were developed, according to each compound’s clinical profile. Adaptive dosing trial simulations were then employed which allowed for estimation of impact of various starting doses and dose modification schema (including degree and number of allowed dose reductions) to be simulated, with future dosing events dictated by the longitudinal ER safety model of myelosuppression. Simulated dosing and subsequent PK profiles were used in the ER models to predict efficacy and safety outcomes for each regimen along with the proportion of patients who remained on investigational treatment in each dose modification schema. Lastly, the clinical utility score of each adaptive dosing regimen was calculated based on weighting of one or more efficacy and safety metrics per clinical relevance and presented in order to guide dosing decisions for the novel compounds. Sensitivity analyses were conducted by varying the weighting of metrics.

Data & Results:

Results: The framework is currently being implemented for novel small molecule compounds in order to determine dosing schema of optimal predicted efficacy and safety, with focus on hematologic adverse events as a driver of dose modifications. This comprehensive framework is guiding recommendations for dose expansion and overall dose optimization in the monotherapy and combination settings. 

Interpretation, Conclusion or Significance:

Significance: This MIDD framework allows for a more holistic and integrated approach for oncology compounds in early to mid-phase development which can be applied to novel monotherapy or combination therapies to inform additional dosing regimens to be evaluated in the trial (relying on simulations of untested dosing regimens) or to support optimal dose selection (relying on clinical data from tested regimens supplemented with simulations). This proposed framework extends previous work to solid tumor indications along with the ability to incorporate longitudinal and logistic regression ER models. Most notably, this approach incorporates model-informed adaptive dosing simulations (dosing interruptions, reductions, and discontinuations), mimicking real world clinical trials and general practice settings where dose modifications frequently occur. This general framework can then be updated and adapted efficiently for use in future compounds and extended to dose-optimization of novel-novel combinations within a variety of indications and exposure-response endpoints.

Disclosures:

All authors are employees and own stock in Pfizer, Inc.

Citations/References:

Additional Information/Authors: None