Dose-Finding Methodologies in Clinical Trials

Dose‐finding represents a critical early stage in clinical development, seeking to identify regimens that achieve therapeutic benefit while minimising harm. Traditional designs, such as the 3+3 escalation‐de‐escalation scheme, rely on fixed cohorts and rigid decision rules, often leading to suboptimal estimation of the maximum tolerated dose (MTD) and prolonged timelines. In contrast, model‐based approaches harness statistical and pharmacological models to inform adaptive dose adjustments in real time. Bayesian frameworks, pharmacokinetic–pharmacodynamic (PKPD) modelling and model‐informed precision dosing (MIPD) offer flexible algorithms that integrate prior knowledge, safety biomarkers and patient covariates to optimise dose selection. Seamless phase I/II designs further streamline the transition from safety to preliminary efficacy assessment by combining escalation and expansion cohorts. Such methodologies have found particular traction in oncology and rare disease settings, where patient numbers are limited and interindividual variability in drug handling and response is pronounced. Advances in computation and data integration now permit incorporation of real‐world evidence, genetic profiles and mechanistic simulations to refine dosing in special populations. Collectively, these innovations promise more efficient trials, improved patient safety and accelerated access to effective therapies on a global scale.

Research from Nature Portfolio

Recent studies have demonstrated the power of adaptive, model‐based designs to improve dose‐finding in oncology. One report introduced a Bayesian hierarchical continual reassessment method that dynamically updates toxicity probability curves using longitudinal biomarker data, resulting in more accurate MTD estimates and reduced patient exposure to subtherapeutic levels. Another investigation presented a seamless phase I/II framework in immuno‐oncology, combining safety and preliminary efficacy endpoints within a unified adaptive algorithm, which shortened trial duration and enhanced identification of the optimal biologic dose. A further study integrated population PK models with real‐world pharmacovigilance data to tailor paediatric dosing regimens, showing improved safety margins and precision in dose recommendations for vulnerable subpopulations.

Dose-Finding Methodologies in Clinical Trials publication trend

The graph below shows the total number of articles in dose-finding methodologies in clinical trials across all publications each year (not limited to Nature Index journals).

Technical terms

Maximum tolerated dose (MTD): The highest dose at which a predefined proportion of subjects experience dose‐limiting toxicity.

Continual Reassessment Method (CRM): A Bayesian adaptive design that updates toxicity probability estimates continuously to guide dose escalation.

Pharmacokinetic–pharmacodynamic (PKPD) modelling: The quantitative linkage between drug exposure over time and the resulting biological effect.

Model‐informed precision dosing (MIPD): A framework that uses individual patient data and mechanistic models to customise dosing regimens.

Seamless phase I/II design: A trial structure that integrates safety and preliminary efficacy assessments within a single adaptive protocol.

References

1. Guidance for statistical design and analysis of toxicological dose–response experiments, based on a comprehensive literature review. Archives of Toxicology (2023).
2. Simulating clinical trials for model-informed precision dosing: using warfarin treatment as a use case. Frontiers in Pharmacology (2023).
3. Computing optimal drug dosing regarding efficacy and safety: the enhanced OptiDose method in NONMEM. Journal of Pharmacokinetics and Pharmacodynamics (2024).