Essentials of Model-informed Drug Development in Oncology

Model-informed drug development (MIDD) is critical to informing the development of safer and more effective new drugs. MIDD can be used to assess drug performance, support regulatory and payor success, characterize diseases, and optimize clinical trial designs.

This blog will focus on best practices for model-informed drug development in oncology.

MIDD maximizes and connects the data collected during non-clinical and clinical development and real-world settings, consisting of individual-level and summary-level information. Learn more about our real-world evidence services here.

It also enables extrapolation to unstudied situations and populations, helping to anticipate the potential risks, mitigate them, and improve the probability of success. Using MIDD can accelerate timelines by bridging across development phases to increase development certainty and leverage existing knowledge and improve cost-effectiveness.

MIDD has become an established approach in the last decade. However, pharmacokinetic/pharmacodynamic (PK/PD) and population PK/PD (PopPK/PD) modeling and simulation (M&S)/pharmacometrics techniques were introduced over 60 years ago. In the 90s, they were largely used experimentally to support drug development programs with limited impact on decision-making. From 2000 to 2010, using PopPK and PK/PD became embedded in drug development and is now critical to many international regulatory guidance documents and frameworks. In addition, global regulatory agencies also encourage integrating MIDD approaches into drug submissions, with increasing recognition of Model Informed Precision Dosing (MIPD) to support precision dosing strategies 

Understanding how an investigational drug’s safety and efficacy profile compares to the standard of care and/or competitor drugs is important medically and commercially. 

Model-based meta-analysis (MBMA) uses highly curated clinical trial data in databases (e.g. Codex) and literature and combined with pharmacometrics models to enable indirect head-to-head comparison, considering the impact of treatment, dosing regimen, patient population, and trial characteristics on responses to medicines. This can support designing and executing pivotal studies. In some cases, MBMA models can even serve as an external control arm. 

Large pharma adopted mechanistic modeling approaches such as physiologically-based pharmacokinetic (PBPK) modeling and simulation, quantitative systems pharmacology (QSP), and semi-mechanistic PK/PD modeling as experimental techniques in the 90s. Nowadays, PBPK modeling to predict first-time-in-human-dosing, drug-drug interactions (DDIs) and explore dosing in special populations, such as pediatric, pregnant, and lactating populations, has become the pharmaceutical industry standard. These models are also frequently used to predict PK in unstudied populations. 

In recent years, the use of QSP has increased. QSP combines computational modeling and experimental data to examine the relationships between a drug, the biological system, and the underlying disease process. By combining mechanistic understanding of molecular pathways with pharmacokinetic and pharmacodynamic data, QSP enables researchers to predict drug effects across different patient populations, explore novel combination therapies, and identify biomarkers for response or toxicity. This approach supports decision-making throughout drug development, from target identification and lead optimization to clinical trial design and dose selection. 

MIDD can be applied to all therapeutic areas and various modalities. This blog highlights its application to oncology therapeutics development. This approach can inform all stages of cancer drug development from early to late clinical development. It allows us to understand an investigational oncology drug’s PK, safety, and efficacy, first in nonclinical species. Then this information can be translated to first-in-human studies where it can inform the dosing strategy (dose, dosing intervals, dose schedule, and dose escalation/de-escalation plan). In addition, PK/PD modeling can help characterize the impact of intrinsic (age, weight, sex, genetic factors, mutations etc.) and extrinsic (food, co-medications, smoking status, laboratory variables etc.) covariates on drug exposure. Concentration-ECG (e.g., C-QTc) analysis using early-stage clinical trial PK-matched QTc data can de-risk the cardiac safety risk or waive the need for a clinical TQT study.  

Oncology drugs are often combined with novel compounds or the standard of care. However, choosing the optimal drug combinations is difficult using conventional clinical trial designs. Emerging mechanistic approaches, such as mechanistic compartmental and QSP modeling, have enabled us to assess downstream biochemical and cellular effects of modulating the target pathway to support developing a drug combination and safety strategy. Lastly, MBMA can be used to help support the optimal trial design, to understand the competitive landscape and in some cases to create an in silico external control arm.  

Ultimately, using MIDD approaches can save time and money while increasing drug efficacy and minimizing risk to the patients.

Jansson-Löfmark R, Fridén M, Badolo L, et al. Translational PK/PD: a retrospective analysis of performance and impact from a drug portfolio. Drug Discov Today. 2025;30(7):104417. doi:10.1016/j.drudis.2025.104417 

Sahasrabudhe V, Nicholas T, Nucci G, Musante CJ, Corrigan B. Impact of Model-Informed Drug Development on Drug Development Cycle Times and Clinical Trial Cost. Clin Pharmacol Ther. 2025;118(2):378-385. doi:10.1002/cpt.3636