�鶹ֱ��

In September 2026, FDA took another significant step toward integrating new approach methodologies, or NAMs, into drug development by issuing a Direct Final Rule on Nonclinical Testing Terminology (the “September 2026 direct final rule”).[1] Read together with FDA’s March 2026 draft guidance on NAM validation, the rule moves the question for NAM developers from whether FDA will consider non-animal data to how a specific NAM earns regulatory trust. As we discussed when FDA released its April 2025 Roadmap to Reducing Animal Testing in Preclinical Safety Studies, the answer has direct implications not only for regulatory strategy, but also for intellectual property and data governance.[2]

The September 2026 direct final rule makes amendments to replace numerous references to “animal” (and, for consistency, “preclinical” and “in vitro”) studies in FDA’s drug and biologics regulations with the broader category of “nonclinical” studies. The new definition expressly covers in vitro, in silico, and in chemico tests, including cell-based assays, organ chips and microphysiological systems, computer modeling, other human or nonhuman biology-based methods such as bioprinting, and animal studies.

In the September 2026 direct final rule, FDA states that the amendments to the rules impose no new testing requirements and are intended principally to remove regulatory terminology that could imply animal testing is the only acceptable source of nonclinical evidence. In particular, the September 2026 direct final rule states that “[t]hese changes result in more consistent and updated terminology that is not unduly focused on animal testing and that encompasses the use of scientifically valid NAMs.”[3] Comments are due December 7, 2026, and the rule is scheduled to become effective February 4, 2027, unless FDA receives significant adverse comments and withdraws it.[4]

FDA’s Guidance on NAM Validation

FDA’s March 2026 draft guidance[5] identifies four related considerations for NAM validation:

• Context of use asks what specific drug-development decision the NAM is intended to support.
• Human biological relevance asks whether the modeled physiology and mechanisms correspond to the human biology relevant to that decision.
• Technical characterization addresses reliability and reproducibility.
• Fit-for-purpose asks whether the NAM performs adequately for its intended regulatory role.

The September 2026 direct final rule echoes this framework, urging developers to consult with FDA and to provide information about a NAM’s technical characterization and biological relevance for particular contexts of use.[6]

Notably, the guidance also states that a NAM does not necessarily need to be formally validated to be considered, and that a fit-for-purpose NAM, even if not validated, may adequately address specific toxicological concerns within a weight-of-evidence assessment. The important point is that evidentiary value depends on the question the model is being asked to answer.

Validation of Models or Studies Is Context-Dependent

The regulatory problem addressed by the FDA guidance and direct final rule is not to identify a universally superior model, but to determine which model, or combination of models, generates evidence appropriate for the decision at hand. In this regard, the direct final rule notes that animal studies may remain important where toxicities arise through complex physiological interactions, including communication among organ systems. At the same time, FDA recognizes that NAMs using human-derived cells may be able to assess additional or more relevant endpoints for clinical drug development.

FDA’s recently published NAM use-case database makes the same point expressly and cautions that inclusion of a method in a prior regulatory review does not amount to general approval of that methodology. Rather, regulatory conclusions remain context-specific and may depend on the particular method, endpoint, product, data package, and application.[7]

Practical Implications

The NAM regulatory framework raises many strategic questions for NAM developers, such as:

• What decision is a pharmaceutical partner or regulator expected to make from its output?
• Which biological features are essential to that decision?
• What evidence demonstrates reliability?
• Where are the model’s limitations?

These questions can shape product development itself. A clearly defined context of use can determine which endpoints should be measured, which comparators are appropriate, what reproducibility studies are needed, and which aspects of a model should be standardized across laboratories. It can also make a platform easier for pharmaceutical partners and investors to evaluate because the developer is offering a defined decision tool rather than a generalized claim of biological realism.

FDA provides concrete ways to seek feedback on these issues. For a specific development program, sponsors may request FDA input on using a NAM, including through a Type D meeting. The guidance indicates that early discussions will focus on considerations specific to the indication, disease, organ, and endpoint. For tools designed to support multiple programs, FDA’s Drug Development Tool Qualification Program and Innovative Science and Technology Approaches for New Drugs (ISTAND) program offer separate pathways.

The most valuable NAM may not be the one that claims to replicate the most biology, but the one that can show most clearly what its model represents, where it is reliable, and why its output can be trusted for a consequential decision. Clarity on these questions is what persuades FDA reviewers, pharmaceutical partners, and investors, and it may help define the boundaries of defensible patent claims. Developers that align context of use, validation evidence, data governance, and IP strategy from the outset will be best positioned as FDA’s framework moves from terminology to practice. We welcome conversations with developers, sponsors, and investors working through these questions.

[1] Nonclinical Testing Terminology, Direct Final Rule, 91 Fed. Reg. 59,988- (Sept. 22, 2026) (effective Feb. 4, 2027, subject to FDA’s direct-final-rule procedures).

[2] /insights/publications/2025/07/navigating-fdas-proposed-guidance-on-ai-and-non-animal-models-safeguarding-inno/

[3] See supra note 1 at 59,993.

[4]See supra note 1.

[5] U.S. Food & Drug Administration, General Considerations for the Use of New Approach Methodologies in Drug Development: Draft Guidance for Industry (Mar. 2026), https://www.fda.gov/media/191589/download; see also Interagency Coordinating Committee on the Validation of Alternative Methods (ICCVAM), Validation, Qualification, and Regulatory Acceptance of New Approach Methodologies (Mar. 2024), https://doi.org/10.22427/NICEATM-2.

[6] Nonclinical Testing Terminology, Direct Final Rule, 91 Fed. Reg. 59,988, 59,992-93 (Sept. 22, 2026) (discussing continued roles for animal studies in complex physiological interactions and the potential for human-derived NAMs to assess additional or more relevant endpoints).

[7] U.S. Food & Drug Administration, New Approach Methodologies (NAMs) Database of Use Case Examples (2026), https://www.fda.gov/drugs/cder-streamlined-nonclinical-studies-and-acceptable-new-approach-methodologies-nams/new-approach-methodologies-nams-database-use-case-examples (noting that regulatory conclusions are context-specific and may depend on the method, endpoint, product, data package, and application).