The Evolving Role of In Vitro–In Vivo Correlation in Model‐Informed Drug Development: A Multi‐Stakeholder Perspective

ABSTRACT

In vitro–in vivo correlation/relationship (IVIVC/R) models such as physiologically based biopharmaceutics modeling (PBBM) are crucial tools that link biopharmaceutical properties to clinical performance. They accelerate development, reduce costly experimental studies and clinical trials, and justify regulatory decisions for drug formulation related questions of interest (QOI). This paper consolidates insights from academia, industry, and service providers, exploring future opportunities, organizational challenges, regulatory perspectives, and competency gaps for further enhanced application in pharmaceutical development and regulatory decision‐making.

1. Introduction

In vitro–in vivo correlation/relationship (IVIVC/R) models such as physiologically based biopharmaceutics modeling (PBBM) are crucial tools that link biopharmaceutical properties to clinical performance. They accelerate development, reduce costly experimental studies and clinical trials, and justify regulatory decisions for drug formulation related questions of interest (QOI) [1, 2]. This paper consolidates insights from academia, industry, and service providers, exploring future opportunities, organizational challenges, regulatory perspectives, and competency gaps for further enhanced application in pharmaceutical development and regulatory decision‐making.

2. Academic Outlook on Biopharmaceutics Research

PBBM provides means for enhanced drug product quality through mechanistic understanding of how drug release is influenced by gastrointestinal (GI) physiology, for example, pH dependence, interactions with colloidal structures and the impact of hydrodynamics. This is especially relevant for Biopharmaceutics Classification System (BCS) class II–IV drugs, that is, drugs with high/poor permeability and/or poor solubility. To further increase confidence in biopharmaceutics evaluations of such drug molecules, additional understanding of the dynamic and highly variable in vivo conditions as well as improved means for experimental assessments are needed.

A key attribute for intestinal absorption following oral administration is intestinal permeability, and expanding our understanding of colonic permeability and absorption mechanisms is essential to refine predictive strategies for assessments of drug development programs that involve drug absorption distal to the ileum. Expanding knowledge with new in vivo reference measurements for this region may be hard to achieve but re‐analysis of existing in vivo data applying PBBM principles, for example, accounting for colonic physiology and regional intra‐luminal conditions, may provide new insights and standards for translations and predictions. Furthermore, emerging technologies, such as intestinal organoids and 3D bioengineered models, could potentially provide means for in vitro assessments with higher biorelevance compared to conventional mono‐layer assays. However, extensive evaluation of such systems remains to confirm their relevance. Finally, the recent surge of high molecular mass drug candidates and new modalities presents new challenges, in terms of solubility, permeability and stability, necessitating the generation of new reference data, the development of new assays and improved predictive models [3]. For instance, advancing techniques such as molecular dynamic simulations and machine learning approaches are expected to play crucial roles in predicting and optimizing the physicochemical properties of new molecules, thereby facilitating the development of more effective and precisely targeted drug therapies.

3. Industry Perspective on Model Informed Drug and Formulation Development

Formulation innovation and development begins early and evolves throughout the drug product lifecycle. An early understanding of critical performance factors is crucial for optimizing formulations for Phase III trials or marketing approval. The first formulation challenge typically arises before First‐in‐Human (FIH) studies, where selecting a formulation requires balancing biopharmaceutics properties with predicted exposure [4]. By connecting quality attributes to in vivo performance, these models ensure consistent product performance in terms of safety and efficacy, in alignment with the target product profile (TPP). Biopharmaceutics inquiries enable high‐impact modeling across the R&D value chain. Tools such as PBBM, compartmental Population PK modeling (PopPK), and deconvolution and convolution‐based IVIVC are instrumental in understanding the formulation‐response relationship (see Figure 1). The choice of approach depends on the specific QOI and available data. In this context, PBBM approaches are increasingly used to complement, extend, or even replace traditional IVIVC/R models, especially when relationships between in vitro and in vivo data are complex, during early stages of development where clinical data are not/less available, or when more mechanistic insights are needed.

FIGURE 1.

The PBBM and PopPK approaches have different advantages and current use in an IVIVC/IVIVR setting. In the future we foresee a gradual merger of these approaches combining their respective strengths.

While these models have been primarily utilized to support oral drug product development, particularly modified‐release (MR) formulations, their underlying principles are readily applicable to a broader range of formulations and routes of administration, including long‐acting injectables [5], transdermal drug delivery, and vaginal rings [6].

These tools support drug developers in decision making and can also be used to facilitate formal IVIVC biowaiver requests for scale‐up and manufacturing changes. Thus, validated Level A IVIVC models offer a critical regulatory and strategic advantage by justifying biowaivers of in vivo bioequivalence studies, thereby accelerating development and reducing costs.

Continuous alignment across functions, that is, DMPK, CMC, Clinical Pharmacology, Pharmacometrics, Regulatory affairs, is key for successful implementation of these approaches to support development programs.

4. Service Provider's Perspective on Model Informed Drug and Formulation Development

For service providers in IVIVC/R, the landscape presents both significant opportunities and challenges. Success requires a fit‐for‐purpose approach with comprehensive expertise across all methodologies, from standard IVIVC and PopPK modeling to sophisticated PBBM. Recent advancements have paved the way for hybrid, “middle‐out” approaches, where mechanistic PBBM methods are integrated with the data‐driven advantages of population modeling to characterize in vitro‐to‐in vivo translational factors [7, 8].

This must be coupled with transparency and platform agnosticism, avoiding the limitations of “black box” solutions. Furthermore, establishing robust internal processes for reproducible research is essential, including clear internal guidance and quality control of modeling scripts and deliverables like reports. Ultimately, the most valuable asset is a team of consultants possessing deep expertise in model‐informed drug development, quantitative approaches and biopharmaceutics—including physicochemical properties, formulation science, absorption mechanisms, and a thorough understanding of regulatory expectations—to effectively guide drug developers in data analysis, experiment design, and trial design.

5. Current and Future Perspectives on the Use of IVIVC/R in Regulatory Decisions

IVIVC/R and mechanistic models are valuable tools for supporting biowaiver claims, reducing in vivo bioavailability (BA) and bioequivalence (BE) studies. Biowaivers are applicable during various stages of drug development, including formulation bridging, post‐approval Chemistry, Manufacturing, and Control (CMC) changes, line extensions, and generic development. They can be obtained through comprehensive risk‐based analyses covering safety, biopharmaceutics, and control strategy, with BCS being a common approach. Where BCS criteria cannot be met, IVIVC provides an alternative pathway for waivers by enabling prediction of BA/BE from in vitro data.

Regulatory agencies encourage new tools that link pharmaceutical quality with clinical performance. The FDA's 1997 IVIVC guidance provides recommendations for developing and evaluating IVIVC predictability, setting dissolution specifications, and using IVIVC as a surrogate for in vivo BE studies for drug approval and post‐approval changes. Similarly, EMA guidelines outline requirements for in vitro dissolution data and pharmacokinetics. The FDA provides specific guidance on using Physiologically Based Pharmacokinetic (PBPK) models for biopharmaceutics applications in oral drug development and manufacturing changes, highlighting model development and validation. In Europe, PBPK model development for biopharmaceutics follows the general PBPK guidance for reporting modeling and simulation, as no dedicated guidance exists (See Supporting Information for regulatory guideline references and references to case examples). Despite the potential of mechanistic modeling, PBPK/PBBM biowaivers still face challenges like insufficient model validation and inadequate in vivo PK data [2, 9]. The recently issued draft ICH M15 on general MIDD principles may facilitate early gap identification and enhance communication with regulators, helping avoid potential rejection of the evidence.

Regulatory agencies have no common consensus on using these methods for generic drugs. However, ongoing efforts are underway. EMA has recently proposed a PBPK‐based framework to support BE biowaivers for complex generics. The FDA has initiated a pilot program to facilitate early interactions for science‐driven topics using model‐integrated evidence (MIE) approaches, to streamline regulatory processes and expedite market access for generics [10]. Early engagement with regulators is critical to overcoming potential barriers, and initiatives like the FDA and EMA's Parallel Scientific Advice pilot program are designed to streamline requirements for complex generic drugs or hybrid medicines.

6. Discussion and Outlook

Model‐based approaches in biopharmaceutics offer numerous opportunities, particularly to overcome the feasibility limitations of traditional BE assessments. These approaches are pivotal in drug repurposing, organ targeting, and developing sophisticated formulations. Leveraging historical drug data, these models offer a powerful method for obtaining BE waivers, particularly in oncology where some drugs often qualify for accelerated access and BE studies cannot be conducted in healthy volunteers. Furthermore, the generic industry can significantly benefit from these approaches by reducing required studies (i.e., from more efficient study design to study waivers) and facilitating market entry. It is vital to extract insights from in vitro data mechanistically, enhancing our collective knowledge and refining models to drive more efficient pharmaceutical development while maintaining high quality standards.

The respective strengths and weaknesses of traditional PopPK convolution techniques versus PBBM approaches can be compared (Figure 1). PBBM, a mechanistic approach, better accounts for complex underlying mechanisms, allowing for more accurate extrapolations. In contrast, PopPK modeling using more parsimonious models excels at capturing inter‐ and intra‐subject variability and performing data‐driven parameter estimation. PopPK models are advantageous for characterizing covariate effects and variability. Although these approaches are often distinct, ongoing advancements are bridging the gap, with improved methods for variability accounting and parameter estimation within PBBM platforms, alongside efforts to incorporate more mechanistic elements into PopPK models for IVIVR purposes [10].

Despite these advancements, this remains an emerging field, necessitating further research to streamline requirements and establish clear regulatory guidelines.

Current perspectives on data input for IVIVC/IVIVR modeling highlight well‐documented practices in the 1997 guidelines, which specify necessary model inputs, including standardized experiments and the required number of distinct formulations. However, current (draft) guidelines fall short on data requirements for PBBM, particularly for aspects like particle size distribution. Data needs vary by development phase; early stages often require limited data packages, while more comprehensive data is necessary in later phases, especially for obtaining waivers.

The best practices for model qualification and validation in this area are continually evolving. Effective models should simulate multiple scenarios and incorporate robust diagnostics to evaluate performance.

Gaps for model‐based approaches in applied biopharmaceutics research reveal several areas for improvement. Effective communication with stakeholders, including CMC formulation teams and pharmacometricians, is crucial for seamless data integration. Continuous organizational efforts are needed to optimize data workflows. Methodological challenges persist in integrating formulation properties and dissolution profiles into models, requiring more mechanistic insights.

This paper highlights the potential of advanced IVIVC/IVIVR models to accelerate drug development and update regulatory processes (concrete examples of applications are available as Supporting Information). Using advanced modeling methods and filling current gaps, the pharmaceutical industry can improve drug delivery systems and accelerate the availability of more effective treatments to patients without compromise.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1: psp470137‐sup‐0001‐Supinfo.docx.