USFDA Guidance: Quantitative Systems Pharmacology (QSP)-Based Dose Selection for Minimum Anticipated Biological Effect Level (MABEL) in First-in-Human (FIH) Trials
- Sharan Murugan
- 1 day ago
- 5 min read
Selecting an appropriate starting dose for first-in-human (FIH) clinical trials is one of the most critical decisions in drug development. An initial dose that is too high may expose participants to unnecessary safety risks, while a dose that is too low may provide limited scientific value and delay clinical development.
To support safer and more informed dose selection, the FDA released the draft guidance "Quantitative Systems Pharmacology (QSP)-Based Dose Selection for Minimum Anticipated Biological Effect Level (MABEL) in First-in-Human (FIH) Trials." The guidance provides recommendations for sponsors on using quantitative systems pharmacology (QSP) models to estimate MABEL doses and support dose selection decisions in early clinical development.
The guidance reflects the FDA’s continued commitment to Model-Informed Drug Development (MIDD) and encourages the integration of mechanistic modeling, biological knowledge, and diverse data sources to improve clinical trial design and regulatory decision-making.
Why MABEL Is Important in First-in-Human Studies
Determining a safe starting dose for FIH studies has traditionally relied on data from animal toxicology studies. Approaches such as the No Observed Adverse Effect Level (NOAEL), Severely Toxic Dose in 10% of rodents (STD10), and Highest Non-Severely Toxic Dose (HNSTD) have historically been used to estimate human starting doses.
While these approaches remain useful for many products, they may have limitations for certain high-risk therapies, particularly those involving novel mechanisms of action, immune activation, or human-specific targets. To address these challenges, the concept of Minimum Anticipated Biological Effect Level (MABEL) was introduced. MABEL represents the dose expected to produce the lowest measurable biological effect in humans while maintaining an appropriate safety margin. The FDA guidance highlights how QSP modeling can strengthen MABEL estimation by integrating biological, pharmacological, and clinical knowledge into a comprehensive quantitative framework.
What Is Quantitative Systems Pharmacology?
Quantitative Systems Pharmacology is a modeling approach that combines biological understanding with mathematical and computational methods to predict how a drug interacts with biological systems. Rather than focusing solely on drug exposure, QSP evaluates the complete sequence of events from drug administration and target engagement to downstream biological responses. By bringing these data together, QSP can help predict the biological effects of a drug in humans and support more informed dose selection decisions.
General Principles for QSP-Based Dose Selection
The FDA outlines several important principles that sponsors should consider when using QSP models to determine MABEL doses.
Integration of Available Data
A fundamental principle of QSP is the use of all relevant information available during drug development.
This may include in vitro studies evaluating receptor binding, target engagement, cytokine release, and cellular activation, as well as pharmacokinetic and pharmacodynamic data from nonclinical studies. Information from disease biology, published literature, and similar compounds may also contribute to model development.
Integrating diverse datasets helps improve the reliability of dose predictions and provides a more comprehensive understanding of drug behavior.
Accounting for Species Differences
For many therapies, especially biologics, important differences exist between animal and human biology.
The guidance recommends that QSP models account for species-specific factors such as receptor expression, physiological characteristics, and pharmacodynamic responses. These considerations help improve the translation of nonclinical findings into clinically meaningful predictions.
Mechanistic Representation of Biology
FDA emphasizes that QSP models should accurately represent the biological pathways involved in drug response. This includes modeling drug absorption, distribution, target binding, signaling pathways, biomarker responses, and other biological processes that contribute to therapeutic effects or adverse reactions.
A mechanistic representation of biology can improve confidence in model predictions and support a more robust understanding of dose-response relationships.
Continuous Learning and Model Refinement
QSP models should not be viewed as static tools. As new information becomes available from nonclinical studies and early clinical trials, sponsors are encouraged to refine and update their models.
This iterative process supports continuous learning and aligns with the principles of Model-Informed Drug Development, where emerging evidence is used to improve future decision-making.
Recommended QSP Modeling Practices
The guidance provides detailed recommendations on how sponsors should develop, evaluate, and document QSP models intended to support FIH dose selection.
Defining the Question of Interest
Sponsors should clearly identify the purpose of the model and the specific questions it is intended to answer.
For example, a QSP model may be used to determine the starting dose, predict an anticipated therapeutic dose range, or support dose-escalation strategies during early clinical development.
Establishing the Context of Use
The intended role of the model should be clearly defined.
Sponsors should explain how model predictions will contribute to decision-making and describe how other sources of evidence will be incorporated alongside model outputs.
A well-defined context of use helps regulators understand the importance of the model in supporting clinical development decisions.
Assessing Model Risk
FDA recommends conducting a model risk assessment based on two key factors:
The degree to which model predictions influence decision-making.
The potential consequences of incorrect decisions.
Understanding model risk helps determine the level of validation and supporting evidence required to establish confidence in model predictions.
Building Biologically Relevant Models
QSP models should reflect current scientific understanding of disease biology and drug mechanisms.
Important components may include target expression, biomarker dynamics, signaling pathways, drug metabolism, transport processes, and tissue-specific drug distribution.
The guidance also encourages sponsors to evaluate alternative model structures when appropriate and transparently document any assumptions or limitations.
Model Verification, Validation, and Application
FDA places significant emphasis on ensuring that QSP models are scientifically credible and technically reliable.
Model Verification
Sponsors should verify that model equations, software implementation, and calculations are functioning correctly. Verification activities may include evaluating expected model behavior, confirming appropriate unit conversions, and ensuring that simulations produce biologically plausible results.
Model Validation
Validation should demonstrate that model predictions adequately reflect observed data.
Whenever possible, sponsors should compare model outputs with findings from nonclinical studies, clinical data, or evidence from similar compounds. Validation helps establish confidence that the model can support important dose-selection decisions.
Application to Dose Selection
Once validated, QSP models can be used to simulate drug responses across different dose levels and predict the biological effects expected in humans. These simulations can support selection of an appropriate starting dose and help identify suitable dose-escalation strategies for FIH studies.
All models contain uncertainty, and FDA recommends that sponsors carefully evaluate its impact on dose selection.
Considerations for Selecting the First-in-Human Dose
Although QSP can provide valuable insights, FDA emphasizes that dose selection should always be based on the totality of available evidence. Sponsors are encouraged to take a cautious approach, particularly when model-based estimates differ substantially from traditional MABEL calculations or when limited clinical experience exists for a target or mechanism of action.
Regulatory Interactions and Submission Expectations
FDA encourages early engagement between sponsors and the Agency when QSP models will play a significant role in dose selection. Discussions during pre-IND meetings and Model-Informed Drug Development (MIDD) paired meetings can help align expectations regarding model development, validation, risk assessment, and clinical implementation.
When submitting an Investigational New Drug (IND) application, sponsors should provide comprehensive documentation describing model structure, assumptions, parameters, validation activities, sensitivity analyses, simulation results, and the rationale supporting the proposed starting dose.
Transparent reporting enables efficient regulatory review and facilitates scientific evaluation of model-based recommendations.
As QSP continues to evolve as a powerful tool for drug development, its integration into MABEL-based dose selection may enhance both participant safety and the efficiency of bringing innovative therapies to patients.
Comments