An AI Foundation Toxicology Model and Framework to Support Waiving a Second Species in Drug Safety Studies

¿De qué trata esta convocatoria?

The topic funds an AI toxicology model and a weight-of-evidence framework that can show when a second animal species is unlikely to add safety information in drug studies. The aim is regulatory acceptance of waivers and fewer animal tests.

¿Qué financia esta convocatoria?

Scope

Regulatory guidelines such as ICH M3(R2) have long required toxicity studies in two species for new small molecule medicines. This is increasingly challenged by ethics, cost, long study times and doubts about human relevance, while new approach methodologies and AI make it possible to reduce reliance on animal models. The topic funds an AI Foundation Toxicology Model, treated as a new approach methodology, together with a weight-of-evidence framework, to predict whether a second species in chronic and sub-chronic small molecule repeat-dose studies would add safety-relevant information.

What proposals should do

• Consolidate and evaluate data from several pharmaceutical companies and other organisations, with priority on general toxicology studies in structured CDISC SEND format and in unstructured reports, and build a database with a common or federated data model that protects sensitive data.
• Select and optimise AI methods that give transparent probabilistic predictions of missed toxicity, organ-specific findings and divergence in no observed adverse effect level, with transparency, interpretability, traceability and data provenance. Methods named include multimodal representation learning, federated and privacy-preserving learning, uncertainty quantification and explainable AI.
• Develop a regulatory strategy and interaction plan, aiming for endorsement by regulators such as the European Medicines Agency and the US Food and Drug Administration by the end of the project, with a longer-term goal beyond its scope of revising ICH M3(R2).
• Build on earlier projects such as BigPicture, eTRANSAFE, eTOX, imSAVAR and VICT3R, and on related international initiatives.

Expected outcomes

• A validated AI Foundation Toxicology Model that predicts when a second species is unlikely to add safety information, with the goal of waiving two-species chronic testing for small molecules and other modalities such as oligonucleotides.
• A standardised, transparent weight-of-evidence framework for industry, regulators and academia.
• Tools, templates and training materials for implementation, including guidance on explainability, provenance, governance, ethical use and alignment with AI requirements.
• Greater industry and regulator confidence in second species waiver applications, supported by calibrated evidence.

The action should contribute to the EU directive on the protection of animals used for scientific purposes and to the replace, reduce and refine principles.

¿Qué se espera de las propuestas?

• Proposals should foster collaboration among pharmaceutical companies, regulatory agencies, academic institutions, SMEs and patient advocacy groups.

¿Cuándo termina el plazo?

8 de octubre de 2026

8 October 2026 for the first stage; the second stage deadline is 21 April 2027. This is a two-stage call.