The pharmaceutical industry is facing a familiar challenge that is becoming increasingly urgent: how to accelerate drug development timelines and reduce reliance on animal models without compromising the scientific rigor required for regulatory decision-making. As expectations evolve within regulatory and scientific communities, the pressure to modernize nonclinical research approaches has never been greater.
Virtual control groups (VCGs) are emerging as one of the most promising innovations. Once considered a conceptual application of historical data, VCGs are now demonstrating real-world potential to adapt how toxicology studies are designed and interpreted. Recent evidence suggests that they are not only feasible, but also capable of preserving the scientific integrity of traditional study designs.
In newly published research from Charles River, VCGs were retrospectively applied to 20 nonclinical toxicology studies to evaluate whether they could replicate the role of concurrent control groups (CCGs). The results showed that across all studies, the use of VCGs led to the same non-observed-adverse-effect level (NOAEL) conclusions as traditional concurrent controls. This complete concordance at the study-decision level shows that reducing reliance on concurrent control animals does not necessarily mean compromising the confidence required for regulatory outcomes.
How virtual control groups work in practice
VCGs leverage high-quality curated historical control data that combines advanced statistical methods and machine learning to generate a matched virtual control group. These datasets are carefully constructed to align with the specific conditions of a given study based on established critical attributes such as, but not limited to species, strain, age, sex, environmental conditions, vehicle and analytical conditions, and instruments.
In practice, this requires a robust data infrastructure that includes a centralized database incorporating and curating raw data from animals and studies to generate VCGs. This curation process is supported by data science pipelines that apply clustering, inferential statistics, and expert validation to identify critical matching criteria. Machine learning algorithms play a key role in identifying subtle associations within the data, enabling selection of control animals that closely mirror the characteristics of the original study population.
The result of these curation processes provides broader and more representative control data sets against which treatment-related effects can be evaluated. In contrast, traditional control groups are limited in size and scope, often providing only a narrow snapshot of biological variability. VCGs expand that perspective, offering scientists a richer framework for interpreting findings, particularly when identifying true drug-related effects and distinguishing those effects from background variability.
Beyond replacement and reduction
While the ability to reduce animal use is a clear and important benefit, defining VCGs solely as a 3Rs (Replacement, Reduction, Refinement) tool undermines its broader impact. VCGs’ underestimated value lies in enabling more informed, data-driven decision-making throughout the drug development process. By providing access to a large, well-characterized pool of control data, VCGs help scientists better understand natural variability and rare background findings. This enhanced context can improve the interpretability of different endpoints, for example in clinical pathology and histopathology, which are areas where small differences in control data can significantly influence identification of findings and conclusions.
Moreover, VCGs open the door to improved study designs. VCG hybrid models combine virtual and concurrent control animals, reducing overall animal numbers while maintaining a flow of new data to sustain the underlying database. Concurrently, this type of design allows for the maintenance of concurrent control animals to assess endpoints that cannot be leveraged by historical data sets. At early-stage discovery, where studies are sometimes conducted without control groups altogether, VCGs can provide a valuable reference point without adding cost or time.
What responsible adoption will require
Despite these promising results, moving VCGs from innovation to standard practice will require thoughtful and collaborative effort across the industry and regulatory agencies. Several key challenges must be addressed to ensure responsible and scalable adoption.
First, data quality and curation practices are crucial. VCGs depend on large, standardized, and continuously updated datasets that accurately capture the complexity of nonclinical studies. This includes not only essential attributes like species and dosing, but also more complex factors such as diet, housing conditions, fasting status, and analytical methods, which can all influence study outcomes and scientific interpretation.
Second, transparency in methodology and validation is essential. The selection criteria used to construct VCGs must be clearly defined and reproducible, with robust qualification frameworks to demonstrate their reliability. Retrospective studies are an important step to qualify those methods, but prospective validation in active studies will be critical for broader acceptance.
Third, clearly defining the scope of studies and also limitation of VCGs will help guide adoption. VCGs can be identified as more appropriate for certain study types, therapeutic areas, or phases of preclinical development than others. Identifying where they add the most value as well as where traditional approaches should remain in place will require ongoing work and shared experience across organizations.
Finally, continuous engagement with regulators will be a key aspect. Regulatory agencies are increasingly open to New Approach Methodologies (NAMs), particularly considering evolving legislative frameworks. However, alignment on expectations, standards, and requirements will be necessary to ensure that VCGs can be confidently integrated into decision-making processes.
Alignment of ethics, efficiency, and scientific rigor
The emergence of VCGs reflects a broader shift in how the industry approaches nonclinical research. Increasingly, the most impactful innovations are those that advance ethics, efficiency, and scientific rigor simultaneously. By curating decades of accumulated data and applying modern analytics, they offer a path toward more predictive and efficient ways of how toxicology studies are conducted. They enable better use of existing knowledge, reduce unnecessary duplication, and provide deeper insight into biological variability, all while maintaining the standards required for regulatory confidence. If done right, VCGs could become more than just an added value to studies. They could serve as a model for the next generation of data-driven toxicology, one in which better science and better ethics go together.
Nataliya Sadekova is Product Owner, Digital Scientific Products, Discovery and Safety Assessment Services at Charles River Laboratories