Introducing the concept of virtual control groups into preclinical toxicology testing

Area of Science:

• Toxicology
• Preclinical research
• Data science

Background:

• Sharing legacy data from in vivo toxicity studies allows analysis of control group variability.
• Historical control data can form a repository for constructing virtual control groups (VCGs).
• VCGs are established in clinical trials but novel for regulatory animal studies.

Purpose of the Study:

• To explore the potential of using virtual control groups (VCGs) in regulatory animal toxicity studies.
• To assess the feasibility of reducing animal use through VCGs.
• To lay the foundation for data sharing and statistical evaluation of VCGs.

Main Methods:

• Collecting and characterizing control group data from subacute (4-week) GLP studies with Wistar rats.
• Sharing control group data among participating companies to investigate cross-company variability.
• Analyzing a set of studies to compare outcomes using VCG data versus real control groups.

Main Results:

• The use of VCGs has the potential to reduce animal use by 25%.
• Prerequisites include large, well-structured control datasets and thorough statistical evaluations.
• Initiatives like eTOX and eTRANSAFE are establishing the foundation for data sharing.

Conclusions:

• VCGs offer a promising approach to reduce animal use in regulatory toxicity studies.
• Successful implementation requires significant advancements in data sharing infrastructure and statistical methodologies.
• Further validation is needed to confirm the impact of VCGs on study outcomes.