P30-74 How advancements in OECD genotoxicity TGs improve chemical safety assessments and challenge retrospective data analysis for NAM based IATA development

Revisions of OECD Test Guidelines (TGs) over time have marked significant advancements in genotoxicity testing, focusing on improved documentation and reporting quality, adaptations of experimental protocols to new scientific and technical knowledge and deeper integration of statistical and biological considerations in data interpretation. Identifying critical changes in OECD TGs revisions for genotoxicity testing is key for (a) understanding variability of data quality within the currently available databases and regulatory assessments and (b) characterising possible uncertainties in performance metrics for Quantitative Structure-Activity Relationships (QSARs) which are built on a large set of historical data. These insights support the devel-opment of robust Integrated Approaches for Testing and Assessment (IATAs) grounded in New Approach Methodologies (NAMs). To this end, we mapped key methodological, analytical, and reporting elements in OECD TGs for both in vitro (N o 471, 473, 476, 487) and in vivo (N o 474, 475, 478, 483, 488) genotoxicity assays widely used in regulatory context. The ongoing updates to TGs are reflected in heterogeneity in data quality within existing databases. Key improvements include more detailed test method reporting and the incorporation of novel techniques such as automated cell scoring methods and kinetochore staining of micronuclei for aneugenicity assessment. In addition, consideration of historical control data and increased number of analysed cells improves statistical power and further strengthens results reliability. Additional information on general toxicity and exposure supports more robust interpretation. These refinements, evident across all reviewed OECD TGs, indicate alignment with the current scientific insights and emphasize the importance of enhanced transparency and reproducibility in both in vitro and in vivo studies. Continuous updates remain critical for advancing internationally harmonized chemical safety assessments. Furthermore, existing databases vary in the level of metadata detail and in the criteria used to determine overall genotoxicity conclusions. For example, we observed that 25% to 57% of compounds with micronucleus data have shown inconsistent genotoxicity calls across the existing public databases such as EURL-ECVAM, ISSMIC and OASIS. While these findings support the value and continued development of the data-rich EFSA-curated genotoxicity database, they also highlight the qualitative inconsistencies coming from the use of large historical datasets which still remain a challenge. This could potentially be resolved by including sources of variability and uncertainty in genotoxicity calls included in current and future common data platforms.