New approach methodologies can reshape preclinical development by combining human-relevant laboratory models, computational predictions, and integrated evidence to assess safety without relying solely on animal studies. Their value depends on whether you select a method suited to the decision, validate its performance, and align its use with regulatory expectations.
For a preclinical team, the shift raises practical questions about reliability, implementation costs, data quality, and regulatory acceptance. You need to know where these methods already support decisions, where animal studies may still be needed, and how to build evidence that reviewers can evaluate. Getting those choices right can improve study planning and reduce dependence on models that do not always reflect human biology.
What Are New Approach Methodologies?
New approach methodologies are non-animal methods, technologies, or combined testing strategies used to evaluate chemical hazards, biological activity, and potential safety risks. They include laboratory-based human cell models, computer predictions, engineered tissues, and methods that combine existing data from related substances.
The term covers more than one device or assay. A new approach methodology could be a simple cell-based test for a specific biological response, a three-dimensional tissue model, or a computational model that predicts toxicity from chemical structure. It could also combine several evidence streams to answer a defined preclinical question.
You should view these tools as decision-focused methods rather than universal substitutes for every animal study. One assay may identify a cellular mechanism, another may estimate exposure, and another may model how several organs respond to a compound. Their combined value comes from connecting each method to a specific safety or development decision.
Why Are Traditional Preclinical Models Under Pressure?
Traditional models face pressure because biological responses observed in animals do not always predict what happens in humans. Differences in metabolism, immune activity, receptor biology, disease mechanisms, and dosing can limit how well animal findings transfer to clinical development.
Animal studies also require substantial time, specialized facilities, and carefully managed study designs. Repeating a study after an unexpected result can delay candidate selection or force your team to investigate whether the finding reflects human risk or species-specific biology. Human-relevant testing can help you examine that uncertainty earlier.
Regulatory modernization adds another reason to reassess established study packages. United States law now permits non-animal and human-biology-based methods to support investigational drug applications, so animal testing is no longer the only allowable route. That change does not automatically remove animal studies from every program, but it gives you more room to justify fit-for-purpose alternatives.
Which New Approach Methodologies Fit Preclinical Work?
The right new approach methodologies depend on the compound, target, expected exposure, organ system, and decision you need to make. A model that works well for local tissue irritation may offer little value for a question involving metabolism across several organs.
In vitro models use cells or tissues outside a living organism to measure biological responses. Three-dimensional tissue models can recreate selected structural and functional features more accurately than flat cell cultures. These models can support focused assessments involving the skin, eye, liver, lung, and other tissues when the test system matches the endpoint.
In silico modeling uses computational methods to estimate chemical activity, exposure, or toxicity. Quantitative Structure-Activity Relationship (QSAR) models compare molecular characteristics with known biological effects, and read-across uses information from related substances to address data gaps. These methods depend on suitable reference data, transparent assumptions, and a clear understanding of where the prediction becomes uncertain.
Organ-on-a-chip systems and other microphysiological systems place human cells within engineered environments designed to reproduce selected tissue functions. High-throughput screening can test many compounds or concentrations across multiple biological targets, and omics methods can measure changes in genes, proteins, or metabolites. Three-dimensional bioprinting may also help your team build more structured tissue models for research and safety assessment.
Are New Approach Methodologies Accepted By Regulators?
Regulators accept some new approach methodologies, but acceptance depends on the agency, product type, endpoint, and intended use. You cannot assume that a method accepted for one chemical category or safety question will replace an animal study in an unrelated drug program.
The United States Food and Drug Administration (FDA) permits alternative methods in drug development and supports the use of qualified tools when they meet applicable scientific and regulatory requirements. The FDA Modernization Act 2.0 removed wording that treated animal testing as the sole acceptable route for certain preclinical evidence. Sponsors may still use animal studies, and reviewers can request additional evidence when a proposed method does not answer the safety question adequately.
The United States Environmental Protection Agency (EPA) maintains a work plan aimed at reducing vertebrate animal testing while protecting human health and the environment. The Organisation for Economic Co-operation and Development (OECD) develops test guidelines and integrated testing strategies that include laboratory, computational, read-across, and other non-animal methods. In Europe, applicable rules require the use of alternatives when scientifically satisfactory methods are available.
Regulatory acceptance is better understood as a use-specific decision than a universal stamp of approval. You need to show what the method measures, why it is relevant to the proposed use, and how its output supports the broader safety assessment. Early communication with the relevant agency can help you identify evidence gaps before they affect submission timelines.
What Do Validation And Qualification Require?
Validation tests whether a method produces reliable, repeatable, and relevant results for its stated purpose. Qualification defines how that validated method can support a particular development or regulatory decision.
A validation package should explain the test system, operating procedure, controls, acceptance criteria, statistical methods, limitations, and sources of variability. Your results should remain consistent across runs, operators, and laboratories when transferability matters. The method should also distinguish known positive and negative responses with enough accuracy for the proposed decision.
Biological relevance matters as much as technical repeatability. A method can produce consistent measurements yet still fail to represent the human process you need to study. You should connect its endpoint to a plausible mechanism, adverse response, or exposure level rather than relying on technical performance alone.
Qualification narrows the claim to what the evidence supports. A liver model may be qualified to identify a defined type of cellular injury under specified conditions, but that does not mean it predicts every form of human liver toxicity. Precise claims build more confidence than broad promises that exceed the model’s tested capabilities.
Can New Approach Methodologies Replace Animal Testing Entirely?
New approach methodologies cannot yet replace every animal study across all preclinical programs. Their strongest current role is often to replace selected tests, reduce the number of animals needed, refine study design, or strengthen a weight-of-evidence assessment.
Complex systemic endpoints remain difficult to reproduce in a single non-animal model. Reproductive and developmental toxicity, long-term effects, immune responses, and interactions among several organs can involve timing, metabolism, circulation, and feedback signals that isolated systems do not fully capture. Linking multiple models can address parts of these questions, but added technical complexity does not guarantee full biological coverage.
Computational methods carry a different set of limits. Their reliability depends on the quality, quantity, and relevance of the data used to build them. A model may perform well for compounds similar to its training data and become less dependable when you apply it to an unfamiliar chemical class.
Organ chips also require careful interpretation. They reproduce selected functions rather than an entire human body, and results can vary with cell source, device design, fluid flow, culture conditions, and measurement methods. You should document those variables and avoid treating technical sophistication as proof of regulatory suitability.
How Can Preclinical Teams Integrate NAMs?
You can integrate new approach methodologies by starting with a defined decision rather than purchasing a technology and searching for a use. Identify the uncertainty in your program, the evidence currently available, and the result that would change candidate selection, dose planning, or safety testing.
Map each proposed method to a specific endpoint and compare it with your current testing strategy. Review its biological relevance, reproducibility, operating range, reference compounds, data requirements, and known limitations. You should also check whether an accepted guideline or agency recommendation covers the intended use.
A staged pilot can reduce implementation risk. Begin with a small set of compounds whose properties and responses are already understood, then compare the new method with historical findings and other evidence streams. Use the results to establish internal controls, performance thresholds, documentation practices, and staff training requirements.
Bring regulatory specialists, toxicologists, computational scientists, laboratory teams, and data managers into planning early. Organ chips, high-throughput systems, and computational toxicology can generate different data formats that become difficult to compare if naming, storage, and reporting standards are inconsistent. A shared data plan makes the evidence easier to review and reuse.
How Should You Build Regulatory Confidence?
Regulatory confidence grows when your claim matches the evidence and your documentation makes the method easy to evaluate. Reviewers need to understand what you tested, why you selected the method, how it performed, and where its limitations begin.
Define the intended use in plain language. State whether the method will screen candidates, investigate a mechanism, prioritize compounds, support dose selection, replace a named study, or contribute to a weight-of-evidence decision. A narrow and testable use case gives you a clearer validation target.
Document deviations, failed runs, excluded data, software versions, model assumptions, and acceptance criteria. Selective reporting can hide variability that later appears during method transfer or regulatory review. Transparent records help another laboratory or reviewer reconstruct how you reached the conclusion.
You should also compare agency expectations before planning a global development program. A method accepted by one authority may require additional supporting evidence elsewhere. Early scientific discussions can clarify whether the proposed package is likely to support the intended submission or needs a bridging study.
What Are New Approach Methodologies?
- Non-animal testing methods
- Include in vitro, in silico, and organ-chip models
- Support human-relevant safety predictions
- Reduce animal testing in preclinical development
Preclinical Development Is Becoming More Decision-Focused
New approach methodologies give you more options for building human-relevant evidence, but no single technology can resolve every preclinical uncertainty. The strongest programs match each method to a defined use, test its reliability, document its limits, and combine it with other evidence where needed. Regulatory changes have opened the door wider, yet scientific performance still determines whether a method can support a decision. If you implement these tools in measured stages, you can improve candidate assessment without replacing established studies before the evidence supports that move. The future of preclinical development will depend less on following one standard study package and more on choosing the best evidence for the question.
References
- United States Food and Drug Administration: New Approach Methodologies
- United States Environmental Protection Agency: New Approach Methods Work Plan
- National Toxicology Program Interagency Center for the Evaluation of Alternative Toxicological Methods
- Organisation for Economic Co-operation and Development: Testing of Chemicals
- European Commission: Animals in Science.
Nirdosh Jagota is Managing Partner at GRQ Biotech Advisors with 30+ years in the biotech industry. A former executive at Amgen, Genentech/Roche, Merck, and Pfizer, he has led >25 NDAs/BLAs/MAAs and hundreds of INDs across global regulatory, quality, and compliance.