Comparative analysis of models for toxicology and carcinogenicity assessment.
| Feature | Traditional 2D cultures | Animal models | Advanced 3D biomaterial-based humanized models |
|---|---|---|---|
| Physiological relevance | Low: lacks tissue architecture, ECM, gradients, and complex cell-cell interactions. | Moderate: has systemic physiology but suffers from critical species-specific differences. | High: precisely engineered ECM, 3D architecture, co-cultures, and physiological gradients mimic human tissue niches. |
| Predictive power for human response | Poor: high false positive/negative rates due to altered cell states. | Variable: often poor, as evidenced by > 90% clinical attrition rate for drugs safe in animals. | Superior: human-derived cells in a human-like microenvironment yield more clinically translatable data on efficacy and toxicity. |
| Immunological relevance | Limited: cannot model complex human immune responses (e.g., TDAR). | Limited: fundamental differences in immune system function and antigen presentation. | High: enables co-culture of human immune and tissue-specific cells to model immunotoxicity, cytokine release, and immunotherapy efficacy. |
| Throughput & cost | High: cheap, scalable, amenable to HTS. | Very low: extremely costly, time-consuming, low-throughput. | Moderate-improving: higher cost than 2D, but throughput is increasing with automation and standardized biomaterial platforms. |
| Ethical considerations | Low concern. | Major concern: significant ethical burden and regulatory push for reduction (3Rs). | Low concern: human-centric, reduces reliance on animal testing. |
| Personalization potential | Limited: primarily uses immortalized cell lines. | None: Uses genetically homogeneous animal cohorts. | High: patient-derived cells [e.g., patient-derived organoids (PDOs)] can be used to create personalized avatars for drug screening. |
| Key limitation | Over-simplification leads to poor predictability. | Species differences lead to poor predictability and ethical issues. | Standardization, characterization, and integration into regulatory workflows are ongoing challenges. |
During the preparation of this work, the authors used Napkin AI for creating the Figures and QuillBot for improving language clarity and grammar. No AI tools were used for scientific writing, data analysis, or interpretation. After using these tools, the authors thoroughly reviewed and edited all content and take full responsibility for the integrity and accuracy of the publication.
SC: Conceptualization, Writing—original draft, Visualization. AD: Methodology, Writing—original draft, Visualization. AS: Methodology, Writing—original draft. PZ: Writing—review & editing. NG: Writing—review & editing. RS: Writing—review & editing. VT: Writing—review & editing. LMA: Writing—review & editing, Supervision. DG: Conceptualization, Writing—review & editing, Supervision. All authors have read and approved the submitted version.
The authors declare no competing financial or non-financial interests related to the content of this manuscript.
Not applicable.
Not applicable.
Not applicable.
Not applicable.
Not applicable.
© The Author(s) 2025.
Open Exploration maintains a neutral stance on jurisdictional claims in published institutional affiliations and maps. All opinions expressed in this article are the personal views of the author(s) and do not represent the stance of the editorial team or the publisher.