Key Takeaways
The EMA's VDS pilot allows NAM data to be reviewed outside a marketing authorization application, reducing concern that assessment of an emerging method could delay or negatively affect an active product authorization.
The pilot reshapes an earlier safe-harbor mechanism that the EMA says saw little to no uptake and is designed to generate practical regulatory experience with NAM data.
Regulatory readiness depends on a defined context of use, biological relevance, performance, robustness, reliability, limitations, and the specific regulatory decision a method is intended to support.
Initial regulatory applicability areas include biologics with limited animal-model relevance, developmental and reproductive toxicity, safety pharmacology, and hepatotoxicity and drug-induced liver injury.
Companies, CROs, method developers, academic laboratories, consortia, and other stakeholders can participate, creating an earlier route for regulatory dialogue around emerging methods.
Why the Original Safe Harbor Was Not Enough
The European Medicines Agency (EMA) has had a regulatory safe harbor for innovative testing approaches for years. According to the agency, however, that pathway saw "little to no uptake." The result was a persistent mismatch: pharmaceutical companies increasingly used new approach methodologies (NAMs) internally, while regulators accumulated relatively little practical experience evaluating the resulting data.1
The original safe-harbor approach allowed data generated with novel methods to accompany data from established approaches in support of a marketing authorization application (MAA). It was intended to help regulators evaluate emerging methods for possible future acceptance, define appropriate contexts of use, understand methodological readiness and limitations, and gain confidence in the resulting data. In practice, much of the evidence generated with NAMs never reached regulators in a form that supported systematic learning. The EMA reports that industry has substantial internal experience applying NAMs in early drug discovery and development, but the data are often absent from clinical trial applications and MAA dossiers because the methods are not part of required regulatory testing. When NAM data are shared, they may be provided largely for information and without enough supporting context to allow regulators to evaluate them systematically.
The voluntary data submission (VDS) pilot launched in September 2026 is the EMA's attempt to change that dynamic. Rather than asking developers to test the regulatory viability of an emerging methodology inside a consequential product submission, the pilot allows NAM data to be reviewed and discussed independently of an MAA. The EMA states explicitly that this separation allows stakeholders to submit data without concern that the assessment could delay or negatively affect an active application.2
The pilot addresses a problem that scientific innovation alone cannot solve. Regulators need exposure to new methods before they can determine where those methods are reliable enough to inform decisions, but developers need a practical way to obtain regulatory scrutiny before staking a product submission on an unfamiliar evidentiary approach.
Reshaping the Safe Harbor
The EMA describes the pilot as an effort to reshape the earlier safe harbor into a more structured VDS framework. The goal is to generate practical experience, identify scientific and operational challenges, and gather evidence that could support a more sustainable voluntary-submission mechanism in the future.
Participation is open to pharmaceutical companies, contract research organizations (CROs), method developers, non-governmental organizations, academic laboratories, consortia, and other stakeholders. The program is not an unrestricted venue for presenting novel technology, however. Submissions must fit predefined contexts of use and Regulatory Applicability Areas (RAAs), and the EMA initially validates and prioritizes applications before inviting selected participants to submit a full briefing package. In some cases, the agency may determine that an Innovation Task Force (ITF) briefing, scientific advice, or qualification advice is a more appropriate route.
Selected applicants then provide a package for scientific assessment and can engage with regulators to clarify the methodology and supporting data. The EMA describes the VDS process as voluntary, nonbinding, and non-decisional, and states that it does not replace or interfere with existing regulatory procedures.
A constructive VDS assessment does not qualify a method for regulatory use or establish that it can replace a conventional study in a future dossier. Participants can instead receive feedback on the regulatory readiness of the method, weaknesses or uncertainties in the supporting evidence, and possible next steps through existing EMA pathways.2
For CROs and method developers, the ability to participate directly is significant. Regulatory discussion does not have to begin only when a pharmaceutical sponsor decides to incorporate a new model into a product dossier. Organizations developing, validating, or generating data with NAMs can bring methods forward for scrutiny within the defined scope of the pilot.
Regulatory Readiness Depends on Context of Use
The EMA's framework makes clear that regulatory acceptance does not attach automatically to a technology label. An organoid, microphysiological system, computational model, or other NAM may be scientifically sophisticated without being suitable for every regulatory purpose. The relevant question is what a specific implementation can reliably contribute to a specific decision.3,4
The EMA formalizes that principle through the context of use (CoU), which defines the intended application of a method and the conditions under which it can inform regulatory decision-making. The CoU should identify the regulatory use or decision context, connect the method to relevant endpoints and biological plausibility, and define its applicability domain and limitations. It may relate to discovery, support for first-in-human development, later clinical trials, or marketing of a medicinal product.
The VDS briefing requirements show what regulators expect developers to provide before they can evaluate whether a method is fit for that purpose. Submissions should explain the relevant regulatory problem, specify whether the NAM is being used independently, in a battery, or as part of a weight-of-evidence (WoE) approach, and describe relevant endpoints and biological mechanisms. Developers are also expected to characterize the method's performance, robustness, biological relevance, reliability, and current or anticipated CoU.
Where applicable, the package should describe reference compounds, portfolio-based validation, external validation, comparisons with in vivo evidence, and measures taken to ensure data reliability. The EMA also asks submitters to address principal advantages and limitations, compatibility with existing guidance, the regulatory gap the method may address, and its potential for broader application.
A method needs more than an attractive biological rationale or a technically impressive output. Regulators need evidence that its performance is sufficiently characterized, its limitations are understood, and its outputs remain interpretable within the specific decision for which it is proposed.
Validation also need not be framed as demonstrating that a technology is universally superior to animal testing. Regulatory acceptance can develop incrementally. A method may become useful for one CoU before another, or contribute as one component of a broader evidentiary package before it can support a decision independently.
Four Regulatory Problems in Focus
The EMA has initially concentrated the pilot on four RAAs: biologics for which conventional animal species are absent or incompletely relevant, developmental and reproductive toxicity (DART), safety pharmacology, and hepatotoxicity and drug-induced liver injury (DILI).
The biologics category illustrates why the challenge is not simply choosing between an animal model and a newer technology. Some biologics have such high target specificity that no pharmacologically relevant animal species exists beyond humans. In other cases, a nonhuman primate (NHP) or another species may show target engagement but reproduce only part of the human pharmacology or downstream biology. The EMA states that NAM-derived evidence may contribute within a scientifically justified WoE approach in such situations.
That evidence may help characterize pharmacodynamics, toxicology, or pharmacokinetics, supplement targeted in vivo studies, and address limitations of available animal models. Under appropriate conditions, the EMA says NAM data may also contribute to first-in-human dose selection and clinical trial initiation and could support reduction, refinement, or replacement of certain NHP studies, including some chronic, reproductive, developmental, or juvenile toxicity studies.
The remaining RAAs apply the same CoU-based logic to other safety questions. In DART assessment, NAM-containing WoE approaches may contribute to characterization of developmental risk and, where the scientific justification and risk-mitigation strategy are adequate, support adaptation or waiving of subsequent studies. The EMA expects the assessment to account for factors including target biology, developmental and reproductive roles, pharmacological risk, patient population, and remaining uncertainty.
For safety pharmacology, NAM data may contribute to screening, hazard identification, mechanistic clarification, and targeted investigation of defined safety concerns. Where sufficiently justified, the evidence may inform evaluation of cardiovascular, central nervous system, or respiratory risk.
For hepatotoxicity and DILI, the EMA identifies potential roles for NAMs in generating mechanistic, functional, and exposure-response information relevant to human liver biology. Such data may contribute to characterization of hepatotoxic liabilities, dose-response relationships, starting and maximum doses, escalation strategies, monitoring, and product labeling.
These four areas demonstrate the regulatory philosophy behind the pilot. The EMA is not asking simply whether a novel platform performs well. It is asking whether the platform can address an identifiable regulatory problem within a defined and sufficiently characterized setting.
The Safe Harbor Is an On-Ramp
The VDS pilot sits alongside other EMA pathways rather than replacing them. Early technologies can be discussed through the ITF, product-development questions can proceed through scientific advice, and sufficiently mature methodologies can move toward qualification. VDS feedback can help determine whether one of those routes should follow.2,4
The EMA's broader NAM framework calls for a clearly defined methodology, endpoints, CoU, evidence of relevance within that context, and demonstrated reliability and robustness. Requirements are also more stringent when a method is intended to support a safety decision than when it is being used for purposes such as primary pharmacology or proof of concept.4
The VDS pilot does not lower the evidentiary bar for replacing established methods. It creates an earlier opportunity to identify what that bar looks like for a particular use, determine where the current data package falls short, and decide whether the methodology is ready for a more formal regulatory pathway.
For developers and service providers, that distinction has practical consequences. A platform intended eventually to support regulatory decisions must be developed not only around technical performance but also around a defensible CoU, known limitations, evidence of reliability, and a clear explanation of the regulatory problem it is designed to address.
A Broader Regulatory Transition
The EMA's initiative is distinct, but regulators elsewhere are confronting many of the same questions. In March 2026, the U.S. Food and Drug Administration (FDA) issued draft guidance establishing a general validation framework and recommendations for NAMs used in drug development. The document emphasizes broadly applicable principles of study design, reporting, and validation rather than endorsing individual technology categories and encourages appropriate NAM use in regulatory submissions.5
Experience within the FDA's Center for Drug Evaluation and Research (CDER) also shows how regulatory adoption can develop through repeated exposure to submitted data. CDER reviewers have described agency experience evaluating NAM-based approaches and incorporating validated methods into regulatory assessment while encouraging sponsors to engage regulators as such tools develop.6
Industry experience reflects similar progress alongside persistent uncertainty. A 2025 pharmaceutical industry survey documented examples of scientifically justified NAM-based biotherapeutic filings accepted by health authorities, while also identifying global harmonization and clinical translatability as continuing concerns. The survey pointed to clearer guidance and stronger regulator-industry collaboration as factors that could support broader adoption.7
Developmental toxicology offers a particularly close parallel to the EMA's diagnosis of the problem. Pharmaceutical companies already use NAMs for developmental toxicity for internal derisking and mechanistic investigation, but a 2025 review found that these methods are still rarely submitted to health authorities for qualification or regulatory decision-making. Greater data sharing, qualification, and regulatory harmonization were identified as important conditions for wider use.8
The common problem is not a shortage of potentially useful methods. It is the need to establish shared expectations around where a methodology is applicable, what evidence demonstrates adequate performance, how uncertainty should be communicated, and when regulators have enough experience to rely more heavily on the resulting data.
Regulatory Learning as Infrastructure
The EMA expects the VDS pilot to remain open through September 2027 and plans to publish a lessons-learned report based on the submissions and assessments. The agency says those findings will inform development of CoU-based acceptance criteria, regulatory-network activities, stakeholder outreach, and international alignment. Individual participants will receive informal feedback on methodological readiness, adequacy of the supporting data, regulatory considerations, and potential next steps.
The pilot's strategic objectives extend beyond individual submissions. The EMA intends to strengthen regulatory understanding of NAM development and application, encourage earlier and more iterative dialogue, support development of CoU-based criteria, promote more consistent assessment across the EU network, and make the route toward qualification and acceptance more predictable.
The earlier safe harbor did not generate enough submissions to build that shared experience. The VDS pilot will test whether a more structured framework can do so. If it does, its most important outcome may be greater predictability about what regulators need to see before NAM-derived evidence can move from promising internal science to a dependable component of regulatory decision-making.
References
1. “Information about the Voluntary Data Submission pilot: Data sharing on New Approach Methodologies under voluntary submission.” European Medicines Agency. 30 Jun. 2026. EMA/149393/2026.
2. “New approach methodologies to replace and reduce animal testing: Voluntary data submission pilot.” European Medicines Agency. 1 Sep. 2026.
3. “Voluntary data submission pilot to advance innovative alternatives to animal testing.” European Medicines Agency. 1 Sep. 2026.
4. “Regulatory acceptance of new approach methodologies (NAMs) to reduce animal use testing.” European Medicines Agency. Updated 1 Sep. 2026.
5. “General Considerations for the Use of New Approach Methodologies in Drug Development.” U.S. Food and Drug Administration, Center for Drug Evaluation and Research. Draft Guidance for Industry. 18 Mar. 2026.
6. Yao, Jia, et al. “FDA/CDER/OND Experience With New Approach Methodologies (NAMs).” International Journal of Toxicology. 45: 136–156 (2026).
7. Shenton, Jacintha, et al. “Opportunities and insights from pharmaceutical companies on the current use of new approach methodologies in nonclinical safety assessment.” Drug Discovery Today. 30: 104328 (2025).
8. Powles-Glover, Nicola, et al. “New approach methodologies for assessing developmental toxicity of pharmaceuticals: Case examples and future directions.” Reproductive Toxicology. 137: 109035 (2025).
