Scientific Literacy 2: New Approach Methodologies
New Approach Methodologies (NAMs) are constantly changing how scientists assess the effectiveness and safety of drugs. Traditionally, animal models have been the standard for preclinical testing. However, they often fall short in accurately predicting human outcomes. Many drugs that turn out to be successful in animal testing fail in humans. Because of this, more researchers are turning to alternative methods such as organs-on-chips, 3D organoids, and computational models, to name a few. These methods provide more accurate results because they use human data and cells, but these promising technologies still have limitations in replicating the complexity of the human body.
One of the main advantages of NAMs is the use of induced pluripotent stem cells (iPSCs) in microfluidic organs-on-chips and 3D organoids. iPSCs are adult somatic cells that have been reprogrammed into a stem-cell-like state, allowing them to differentiate into various cell types. This makes patient-specific models of human tissues possible for personalized care. Organs-on-chips are small devices with channels that simulate blood flow and tissue structures, while organoids are clusters of cells that resemble miniature organs. Both systems create a more accurate environment than traditional animal models because they use human cells, enabling researchers to replicate certain functional features of these complex cells.
More specifically, the Nature article explains how researchers can generate iPSCs from patients and use them to create organoids for modeling specific diseases. This enables “clinical trials in a dish,” where drugs are tested directly on human-like systems before they are used on actual patients (Kwon, 2026). Similarly, organs-on-chips, like liver chips, can imitate how drugs move through human tissues and accurately predict toxicity. In one case, a liver-on-a-chip model correctly identified drugs that cause liver injury, and even detected harmful compounds that previously passed animal testing (Kwon, 2026). An article by Barua et al (2025) emphasizes that microfluidic organ-on-chip technologies combine IPSC-derived cells with controlled environments, such as fluid flow and mechanical forces, to closely mimic real human physiology. This level of control allows researchers to observe cell behavior under conditions that resemble those in the human body, improving the reliability of drug testing results.
In addition to physical models, computational approaches and generative AI are becoming increasingly important in replacing animal testing. These systems predict how drugs will behave using large datasets from human, animal, and laboratory studies. For example, computational models can assess whether a chemical causes skin sensitization, a common safety test traditionally conducted on animals. One model was developed using data from hundreds of chemicals and was able to accurately predict allergic reactions through pattern recognition (Kwon, 2026).
Generative AI systems go a step further by simulating entire biological responses. The Nature article describes a model called AnimalGAN, which was trained on data from thousands of laboratory animals. It can generate virtual test subjects and predict outcomes like liver toxicity. When used in simulated experiments, this model successfully ranked drugs based on their potential to cause liver damage (Kwon, 2026). The AIP article supports this by explaining that AI-driven platforms integrate diverse datasets to model complex biological interactions, making them powerful tools for toxicology predictions (Barua et al., 2025). These computational methods are especially valuable alternatives because they are efficient, less costly, and more ethical.
Despite the advances, NAMs still face significant biological and technical limitations. A major issue is that many models are “reductionist,” meaning they focus on specific cell types or systems rather than the entire organism. For instance, a kidney-on-a-chip might only include one type of kidney cell, even though a real kidney contains many different cell types that interact (Kwon, 2026). This makes it difficult to fully replicate how drugs affect whole systems.
Another limitation is the difficulty in modeling interactions between different organ systems. In the human body, organs communicate through networks such as the nervous system. These interactions are highly complex and not yet fully understood by current NAMs. Processes like aging, immune responses, and hormonal regulation are hard to recreate in a laboratory setting. The Nature article notes that whole-organ interactions and tissue aging remain major challenges, so animal studies are still needed in some cases (Kwon, 2026).
Technical challenges also exist because, while AI models are powerful, they are only as good as the data they are trained on. This can lead to biases or limits in their predictive accuracy.
In summary, New Approach Methodologies are transforming drug testing by providing more human-relevant and ethically responsible alternatives to animal testing. Technologies like IPSC-derived organoids, organs-on-chips, and AI-based computational systems allow for more accurate predictions of human responses. However, their current limitations, especially in modeling whole-body processes, mean they cannot yet fully replace animal testing. With ongoing research, scientists can unlock the full potential of these methods and move toward a future where animal testing is no longer necessary.
References
Barua R., Das D., & Biswas N. (2025). Revolutionizing drug evaluation system with organ-on-a-chip and artificial intelligence: A critical review. Biomicrofluidics, 19(6), Page1. 10.1063/5.0268362
Kwon, D. (2026, February 25). The age of animal experiments is waning. Where will science go next?. Publication_Title, https://www.nature.com/articles/d41586-026-00563-3
An L., Liu Y., & Liu Y. (2025). Organ-on-a-Chip Applications in Microfluidic Platforms. Micromachines, 16(2), 201. 10.3390/mi16020201