When Clinical Pharmacology Cannot Wait for Complete Data: Reflections Following the 2026 ACCP Annual Meeting

Abstract

Another American College of Clinical Pharmacology (ACCP) Annual Meeting is successfully in the books. As always, the meeting offered a wide range of stimulating symposia and sessions as well as many thoughtful conversations that stayed with many of us well after the rooms emptied. Among them, Symposium 1, “Where Are We Now in Pediatric Cancer Drug Development: 5 Years of the RACE for Children Act,” brought together perspectives from FDA, industry, and pediatric oncology on a question that is both technically demanding and deeply human: how do we move promising therapies toward children with cancer when the evidence can never be as abundant as we would like?

That question becomes real when we picture the child in front of us: a child with a serious, life-threatening rare disease and no approved treatment. A therapy may have compelling biology and encouraging adult data, yet the appropriate pediatric dose, exposure target, benefit–risk profile, and durability of response remain incompletely characterized. In that moment, clinical pharmacologists are asked to confront a difficult but increasingly common question: should development wait for conventional evidentiary completeness, or should the available evidence be integrated quantitatively to support a timely, proportionate, and ethically defensible decision?

This question should not be construed as an invitation to lower evidentiary standards. Rather, it requires standards that are fit for purpose when large dose-ranging studies, conventional randomized trials, or extensive pediatric efficacy programs are infeasible, unacceptably slow, or ethically difficult to conduct. In pediatric rare diseases, the relevant goal is not “perfect” evidence. It is an evidence package that is mechanistically coherent, quantitatively rigorous, transparent about uncertainty, and appropriate to the severity of the condition, the unmet need, and the feasibility of obtaining additional data.

The ACCP Annual Meeting brought renewed focus to this central challenge in drug development. Discussions spanning pediatric oncology, pediatric extrapolation, model-informed dose selection, and rare-disease development underscored an expanding responsibility for clinical pharmacology: to transform sparse, fragmented, and heterogeneous evidence into decisions that are scientifically credible and clinically useful.

For readers of CPDD, this is not a peripheral issue. It lies at the intersection of pharmacokinetics, pharmacodynamics, translational science, quantitative systems pharmacology, regulatory science, and therapeutic decision-making. Clinical pharmacology is increasingly expected not merely to describe variability, but to explain it, predict its consequences, and guide action despite residual uncertainty.

Pediatric drug development has long been constrained by small populations, developmental heterogeneity, ethical limits on research participation, and delayed integration of pediatric plans into adult programs. In oncology, the Research to Accelerate Cures and Equity for Children Act¹ has further shifted the paradigm by enabling the FDA to require pediatric investigation of certain molecularly targeted oncology products when the molecular target is substantially relevant to the growth or progression of a pediatric cancer. This approach moves pediatric evaluation beyond a narrow focus on whether the adult tumor type occurs in children and instead emphasizes underlying disease biology.²

This move is scientifically appropriate, but it also sharpens the practical challenge. Relevant molecular alterations may occur in only a small number of children across multiple tumor types. A conventional sequence of adult development, followed by pediatric dose-finding, followed by a large randomized pediatric efficacy trial, is frequently unrealistic. In some settings, it may be neither operationally feasible nor ethically appropriate.

However, the answer is not to abandon rigor. Rather, it is important to be clear about what rigor looks like when conventional development pathways are not realistic: careful integration of evidence, transparent assumptions, and a plan to test what remains uncertain.

This is especially important when sponsors and regulators must decide whether adult data are informative for children; whether exposure matching is appropriate; whether a biomarker can support dose selection or early efficacy inference; and what additional pediatric information is necessary before initial approval, labeling, or broader use. Each of these questions is fundamentally a clinical pharmacology question.

In pediatric rare diseases, clinical pharmacology should be considered a decision science. The task is not simply to collect pharmacokinetic data in children, nor to fit models after the development program is substantially complete. It is to prospectively integrate evidence from multiple sources and use that evidence to guide dose selection, study design, development sequencing, and post-approval verification.

Recent work on model-informed drug development (MIDD) in pediatric rare diseases emphasizes the value of integrating complementary quantitative approaches across the drug-development lifecycle, rather than applying individual models in isolation.³ Population PK, PBPK, exposure-response analysis, disease models, and real-world data can collectively make limited pediatric observations more informative, while also clarifying the assumptions that must hold for an extrapolation strategy to be credible.³

Although a model cannot replace evidence, it can allow the development team to state clearly what the evidence supports, what it does not support, and what information would materially change the decision.

Pediatric dose selection is sometimes described in overly simple terms: allometric scaling, exposure matching, or adaptation of an adult dose. Each of these can be useful; however, none is universally sufficient.

Body size matters, but body size is not development. Weight-based or body-surface-area–based dose scaling may be reasonable in some settings, particularly for older children when clearance mechanisms and disease biology are like those of adults. However, it may be inadequate when maturation of renal function, hepatic metabolism, plasma proteins, transporters, gastrointestinal physiology, or the target pathway materially affects PK or PD.

For this reason, simple mg/kg scaling should be approached as a starting hypothesis that requires evaluation against known developmental physiology, available pediatric data, and the clinical consequences of exposure misclassification.

When these assumptions are weak, exposure matching alone may provide an insufficient basis for dose selection. The appropriate question is not whether pediatric exposure resembles adult exposure in a statistical sense. It is whether the predicted exposure is likely to produce the intended pharmacological effect and acceptable safety profile in the pediatric population of interest.

The International Council for Harmonisation's E11A guideline provides a contemporary framework for pediatric extrapolation.⁴ It emphasizes that extrapolation is an iterative process: developers should assess the information available from a reference population, identify the gaps relevant to efficacy, safety, and dosing, and generate only the additional data needed to address those gaps. The guideline identifies modeling and simulation as essential tools for addressing data limitations, ethical constraints, and logistical challenges in pediatric drug development.

The growing role of modeling in pediatric rare diseases makes transparency non-negotiable. A model-informed recommendation should not be presented as a black-box output or as a single point estimate that obscures uncertainty. It should clearly describe the biological and statistical assumptions that underpin the recommendation.

For a PBPK model, this includes the evidence supporting drug-specific inputs, system parameters, maturation functions, disease-state assumptions, and the predicted effects of food, formulation, organ impairment, and concomitant medications. For a population PK model, it includes the adequacy of the sampling design, the handling of below-quantification data, covariate model development, residual-error assumptions, model diagnostics, and uncertainty in individual exposure predictions. For exposure–response models, it includes the biological plausibility of the relationship, potential confounding, adequacy of exposure variability, endpoint reliability, and the distinction between association and causation.

Sensitivity analyses are particularly important. If a pediatric dosing recommendation depends materially on uncertain enzyme ontogeny, bioavailability, target-mediated disposition, a disease-specific clearance pathway, or an assumed exposure–response slope, this dependency should be visible. The development team should ask not simply, “What dose does the model predict?” but also, “Which assumptions most affect that prediction, and what data would reduce the uncertainty enough to alter the decision?” This distinction is critical in rare diseases, where observed data may be sparse, and the apparent precision of a model can exceed the true certainty of the underlying biology.

The credibility of model-informed development does not come from pretending uncertainty has disappeared but rather from acknowledging uncertainty, understanding its consequences, and managing it with discipline.

Rare diseases create an evidentiary environment in which no single study or analysis may be decisive. The challenge is to develop a coherent body of evidence in which independent components converge: disease biology, genotype–phenotype relationships, natural history, nonclinical pharmacology, target engagement, pharmacodynamic biomarkers, exposure–response relationships, clinical outcomes, and safety data.

The FDA's Rare Disease Evidence Principles⁵ reflect this emerging regulatory perspective for a narrowly defined set of programs involving very small, genetically characterized patient populations with severe disease and limited or no adequate treatment alternatives. In appropriate cases, the framework contemplates that substantial evidence of effectiveness may be established through one adequate and well-controlled study supported by robust confirmatory evidence. The latter may include strong mechanistic evidence, relevant nonclinical models, clinical pharmacodynamic findings, natural-history data, expanded-access experience, or case reports.⁵

This framework should not be generalized indiscriminately. It is neither a waiver of evidentiary standards nor a routine substitute for robust controlled evidence. Rather, it recognizes that the strength of an evidence package can arise from the convergence of multiple, independent, biologically plausible, and prospectively specified sources of information.

Clinical pharmacology is central to this cumulative-evidence model. Pharmacokinetic characterization establishes whether the intended exposure has been achieved. Pharmacodynamic data establish whether the drug engages its target or modifies a biologically relevant pathway. Exposure–response analyses assess whether higher or lower exposures are associated with meaningful differences in biomarker response, clinical benefit, or toxicity. Quantitative disease models provide context for interpreting longitudinal change. Together, these data can increase confidence that an observed clinical effect is pharmacologically credible rather than incidental.

Natural-history data are likewise indispensable. Well-designed natural-history studies can define the pace and variability of disease progression, identify clinically meaningful changes, characterize prognostic factors, and support endpoint development. They should be developed prospectively, with transparent data provenance, endpoint harmonization, attention to missingness, and a clear understanding of comparability with the treated population.

Pediatric rare diseases combine two independent sources of uncertainty: limited disease-specific clinical evidence and developmental changes in pharmacology. This is why they should not be approached simply as adult rare diseases in smaller patients.

A therapy may have a well-established mechanism but uncertain pediatric pharmacokinetics. Alternatively, pediatric exposure may be adequately predicted, while the relationship among exposure, target engagement, biomarker response, clinical outcome, and toxicity remains uncertain. In some programs, the main question will be dose selection. In others, it will be the relevance of the clinical endpoint, the validity of an external comparator, the time course of benefit, or the generalizability of data from an older pediatric cohort to infants or young children.

Recent analyses of pediatric rare-disease development have highlighted the need for earlier quantitative planning, improved disease-progression knowledge, more systematic use of model-informed approaches, and stronger collaboration across sponsors, investigators, patient communities, and regulators. These needs are not merely operational. They determine whether limited clinical data can be interpreted with the level of confidence required to support a therapeutic decision.

Clinical Pharmacology in Drug Development can help advance the field by creating space for rigorous, candid discussion of how decisions are made under evidence constraints. We need a more disciplined account of when extrapolation is credible, when models meaningfully reduce uncertainty, when conventional trials remain essential, and how residual uncertainty should be communicated and managed.

Such work would be particularly valuable if it places the clinical pharmacologist at the center of the development team's decision architecture. A clinical pharmacologist has the ability to connect pharmacology, physiology, disease biology, clinical measurement, and regulatory context in ways that make drug development more informative and more rational.

The ACCP meeting brought a familiar reality into sharper focus: pediatric rare-disease development cannot be reduced to a false choice between waiting for complete conventional evidence and acting too soon. The appropriate path is neither therapeutic nihilism nor uncritical reliance on models. It is a disciplined, transparent, model-informed approach in which mechanistic understanding, quantitative analyses, empirical observations, and prospective verification are integrated from the outset.

In pediatric rare diseases, we should ask whether we have used every available data source and every appropriate quantitative tool to ensure that the data we can realistically obtain are maximally informative. We should ask whether the assumptions underlying extrapolation are explicit, whether uncertainty has been characterized honestly, and whether a credible plan exists to confirm or revise the initial decision.

That is not a relaxation of scientific standards. It is clinical pharmacology at its most useful and rigorous: making the best possible decision with the evidence available today, while deliberately building the evidence needed for a better decision tomorrow.