02027nas a2200313   4500000000100000008004100001260001500042653000800057653002300065653003200088653002100120653003200141653001600173653001400189653001500203653001600218653002200234653001500256100001600271700002300287700001800310700001800328245011300346856007200459300001100531490000800542520114900550022001401699       2026                            d  c2026-06-0110a3Rs10aAdvanced analytics10aAnimal Testing Alternatives10aDrug development10aNew approach methods (NAMs)10aNonclinical10aReduction10aRefinement10areplacement10aSafety assessment10aToxicology1 aLaura Lotfi1 aT. William O'Neill1 aAngela Wilcox1 aSteven Bulera00aVirtual control groups in nonclinical research: A Re-analysis of 20 studies using virtual control group data  uhttps://www.sciencedirect.com/science/article/pii/S0273230026000504  a1060770 v1683 aAnimal models are crucial in biomedical research, particularly in pharmaceutical discovery and safety testing. Recent legislative updates and regulatory shifts, such as the FDA Modernization Act 2.0 and European Parliament Resolution 2021/2784 (RSP), have opened new avenues for innovative nonclinical research. Among these innovations is the use of Virtual Control Groups (VCGs), a method which can significantly reduce animal usage and improve experimental designs. VCGs utilize historical control data and machine learning algorithms to replicate traditional control groups' statistical power and address preanalytical and analytical variations. We applied VCG data retrospectively to 20 pilot studies. The VCG and concurrent control group (CCG) data were compared to assess statistical alignment as well as biological relevance. The findings highlight the effectiveness of VCGs in maintaining toxicological study integrity and underscore the importance of specific selection criteria in ensuring accurate toxicological outcomes. This work supports the potential of VCGs in advancing the 3Rs and improving nonclinical research methodologies.  a0273-2300