Key Takeaways
- Over 90 % of drugs fail in clinical trials, largely because preclinical models lack physiological relevance to humans.
- New Approach Methodologies (NAMs) combine stem‑cell‑derived cultures, organoids, organ‑on‑chip systems, and artificial intelligence (AI) to create more human‑relevant, predictive tools.
- AI‑driven in silico models excel at early‑stage target identification, molecular design, and safety screening, while experimental NAMs provide mechanistic validation and translational grounding.
- NAMs can capture human biological diversity—genetics, age, sex, environmental exposures—through donor‑varied iPSC biobanks and “cell/organoid villages,” improving efficacy and safety predictions.
- Widespread adoption hinges on improving biological fidelity, data quality, standardization, and building regulatory confidence through transparent validation frameworks and global collaboration.
- Realizing NAMs’ full potential requires evolving training programs, multidisciplinary workforces, and public‑private partnerships to democratize access and avoid concentration of expertise in well‑resourced institutions.
Introduction: The Problem with Current Preclinical Models
Developing a new medicine remains a long, costly, and risky endeavor. Despite decades of scientific progress and billions of dollars invested in research and development, roughly nine out of ten drugs that enter clinical trials ultimately fail. A major contributor to this high attrition rate is the limited physiological relevance of traditional preclinical models, which often rely on animal systems that cannot fully recapitulate human biology. Consequently, researchers are shifting focus from refining animal models toward constructing more predictive, human‑relevant platforms.
What Are New Approach Methodologies (NAMs)?
New Approach Methodologies (NAMs) represent an evolving collection of technologies designed to improve how we study disease, predict drug safety and efficacy, and assess other regulated products. Rather than depending on a single breakthrough, NAMs integrate advances from stem‑cell biology, organoid culture, organ‑on‑chip microfluidics, and artificial intelligence (AI). According to Dr. Joseph C. Wu of the Stanford Cardiovascular Institute, the emerging paradigm is not replacement but integration: computational approaches enable rapid prioritization and design, while experimental NAMs provide mechanistic validation and translational grounding, together supporting a more predictive, scalable, and human‑centric drug discovery pipeline.
Components of the NAM Toolkit
The NAM toolkit includes several complementary technologies. Stem‑cell‑derived models, especially those generated from induced pluripotent stem cells (iPSCs), allow scientists to grow specific human cell types—such as cardiac, neuronal, or hepatic cells—in the lab. Organoids take this further by enabling these cells to self‑organize into three‑dimensional miniature tissues that mimic key aspects of organ structure and function. Organ‑on‑chip and microphysiological systems recreate the physical and mechanical microenvironment of living tissues, letting researchers study cellular responses under more physiologically realistic conditions. Importantly, these experimental systems are not meant to replace one another; they form an interconnected toolkit where the strengths of each approach offset the limitations of the others.
Role of AI and In Silico Modeling
Artificial intelligence plays a transformative role, especially in the earliest phases of drug discovery. Computational models can swiftly screen vast chemical libraries, identify promising drug targets, and predict potential toxicities far faster than traditional laboratory assays. Dr. Wu notes that in silico approaches deliver the greatest value in target identification, molecular design, and safety prediction, accelerating timelines and reducing costs by replacing physical throughput with computational exploration. AI also expands access to underexplored chemical spaces, such as the hidden diversity within natural products, and can mine existing clinical and real‑world datasets to uncover population‑specific differences in drug response that were missed in conventional trials. However, AI’s predictions are constrained by data quality, representativeness, and standardization, necessitating experimental validation for mechanistic resolution.
Capturing the Diversity of Human Biology
One of the most significant advantages of NAMs is their ability to reflect the genetic, biological, and environmental diversity found across human populations. Traditional preclinical models often fail to account for variations that influence individual drug responses—such as polymorphisms in drug‑metabolizing enzymes, sex‑based differences, or environmental exposures. Human‑derived NAMs address this by incorporating cells from diverse donors and pairing them with AI‑driven analyses capable of scaling variation assessment. Wu envisions integrating in silico models with stem‑cell‑ and organoid‑based NAMs to systematically capture diversity across genetic background, age, sex, and exposures, enabling both retrospective and prospective evaluation of drug efficacy and safety. Examples include combinatorial NAM platforms for drug‑induced cardiotoxicity that assess arrhythmia, fibrosis, and mitochondrial dysfunction while accounting for inter‑individual differences, and large‑scale iPSC biobanks or “cell/organoid villages” that simulate entire patient populations before clinical trials begin.
Building Confidence and Overcoming Challenges
Despite their promise, NAMs are not flawless. Many experimental systems still require improvements in maturity and reproducibility, and AI models remain heavily dependent on the quality and diversity of training data. Regulators demand robust evidence that these approaches can consistently deliver reliable, clinically meaningful results. Wu identifies three main bottlenecks to translational adoption: biological fidelity, data quality, and regulatory confidence. Overcoming these challenges will depend on demonstrating robustness and transparency across academia, industry, clinicians, the public, regulatory agencies, and patient advocacy groups, without stalling innovation. The field should focus on developing standardized validation frameworks, integrating complementary technologies, and ensuring FAIR (findable, accessible, interoperable, reusable) data sharing. Automation, scalability, and global collaboration are also crucial for moving NAMs beyond specialist settings toward broader adoption.
A New Generation of Drug Discovery: Training and Collaboration
Realizing NAMs’ full impact will require new approaches to training researchers and stronger collaboration among academia, industry, and regulatory agencies. Current academic and industry training models, while solid, are not yet fully aligned with a future where AI, organoids, and regulatory science converge. Wu advocates evolving toward a new kind of drug discovery science that embraces multidisciplinary expertise. NAMs can democratize drug discovery if deliberately designed with accessibility, standardization, and global collaboration in mind. However, there is a risk that these technologies become concentrated in well‑resourced institutions due to their technical complexity, infrastructure demands, and data requirements. To mitigate this, Wu calls for building a multidisciplinary workforce embedded in collaborative ecosystems, aligning technical innovation with regulatory insight and societal responsibility through public‑private partnership‑driven frameworks. Such efforts can empower a new generation of scientists capable of translating NAMs‑driven advances into practical, trustworthy, and globally accessible solutions.
Conclusion: Toward More Predictive, Human‑Centric Medicine
The shift from animal‑centric to human‑relevant preclinical research is gaining momentum through the adoption of New Approach Methodologies. By integrating stem‑cell‑derived cultures, organoids, organ‑on‑chip systems, and AI‑driven in silico modeling, NAMs offer a more predictive, scalable, and diverse approach to drug discovery. While challenges related to biological fidelity, data quality, and regulatory acceptance remain, coordinated efforts to standardize validation, share data openly, and foster global collaboration can build the confidence needed for widespread implementation. Ultimately, NAMs have the potential to reduce clinical‑trial failure rates, bring safer and more effective medicines to patients faster, and make the drug discovery process more inclusive and equitable.