Scientific and Regulatory Policy Committee Points to Consider: Virtual Control Groups and Their Implications for...

Replacement of concurrent control groups (CCG) with virtual control groups (VCG) has been advocated as a strategy to reduce animal use in nonclinical toxicity testing. This "Points to Consider" paper examines VCG utility as it impacts toxicologic pathology endpoints and identifies the following points. (1) The CCG remains the gold standard, as it best controls for extrinsic and intrinsic study variables in discriminating subtle effects. (2) Pure VCG are unlikely to provide the same discernment since historical control data (HCD) from which VCG are assembled do not share all biological, environmental, husbandry, and procedural attributes of animals in the current study. (3) Clinical pathology endpoints are particularly vulnerable to VCG distortion given their dynamic range and sensitivity to preanalytical variables, with proof-of-concept studies demonstrating nonreproducibility rates for data interpretation of 31% to 49%. (4) Hybrid control groups combining reduced CCG with VCG are the most rational implementation strategy, balancing 3Rs (reduction, refinement, replacement) goals with scientific rigor and ongoing HCD refreshment. (5) The HCD pools require major upgrades, including comprehensive individual animal metadata, particularly for preanalytical and analytical variables, and (6) access to archived tissues as whole-slide scanned images, permitting reevaluation by the study pathologist using consistent nomenclature and grading criteria. (7) No major regulatory authority currently accepts VCG as a replacement for CCG in Good Laboratory Practice studies, and (8) proprietary data and data-sharing barriers to building sufficiently large multi-institutional HCD databases remain unresolved. (9) The VCG feasibility is highly species-dependent. (10) Multiple other new approach methodologies and modified study designs can support reduced animal use when VCG are not an optimal approach.