Elba Raimúndez 1,5, Ana Victoria Ponce-Bobadilla 2,5, Sven Stodtmann 3,5, Jane Knöchel 4,5
1 Sanofi (Frankfurt, Germany), 2 Certara (Radnor, USA), 3 Abbvie Deutschland GmbH & Co. KG (Ludwigshafen, Germany), 4 Department of Drug Design and Pharmacology, University of Copenhagen (Copenhagen, Denmark), 5 as part of the ISoP AI/ML SIG Steering Committee (, )
Introduction
Artificial intelligence and machine learning (AI/ML) approaches are increasingly being applied across pharmacometrics, including QSP, PBPK and population PK/PD. The rapid uptake of AI/ML has highlighted critical gaps in reproducibility, benchmarking, interpretability, and alignment with regulatory and clinical decision-making. These challenges align with recent regulatory emphasis on the need for transparent AI model development practices and rigorous, risk-based performance assessment frameworks [1]. The scientific value and translational credibility of AI/ML models will yield into meaningful real-world impact when the community agrees on general methodological standards, adopts transparent and reproducible evaluation practices, and engages in open ongoing discourse about model limitations, assumptions, and domains of appropriate application. The new Special Interest Group (SIG) in AI/ML under the International Society of Pharmacometrics (ISoP) was created to address these needs.
Objectives
– To promote rigorous and reproducible AI/ML methodologies in pharmacometrics
– To create and maintain benchmarking datasets to enable a transparent evaluation
– To develop and disseminate best practices for model development, validation and evaluation
– To foster a multidisciplinary community forum for scientific discussion
Methods
The ISoP AI/ML SIG was designed as an action-oriented, community-driven initiative to accelerate rigorous and impactful AI/ML use in pharmacometrics. A multidisciplinary steering committee brings together expertise across industry, academia and clinical research organizations. Cross-organizational partnerships with other ISoP Special Interest Groups facilitate knowledge sharing, ensure alignment, and create synergies across the field. Activities are organized into coordinated workstreams covering: (i) technical innovation and best practices, (ii) education, outreach, and knowledge sharing, and (iii) regulatory requirements and readiness for decision-making.
Results
The AI/ML SIG has established a formal charter defining its mission, leadership structure, and engagement opportunities. A key initiative for 2026 is the development of curated benchmark datasets, with the goal of anchoring this effort in the ISoP journal by 2027. Activities include scientific sessions at major conferences, satellite events including a dedicated workshop at PAGE 2026, regular educational webinars, and focused working groups — creating multiple entry points for participation as contributors, educators, reviewers, dataset providers, or leaders shaping SIG initiatives.
Strong international interest from across the pharmacometrics community highlights the value of a dedicated forum centered on methodological rigor, transparency, and real-world impact rather than tool promotion [2]. Regulatory perspectives are embedded throughout to ensure alignment with decision credibility and patient safety.
Conclusions
By emphasizing benchmarking, reproducibility, and transparent evaluation, the SIG aims to accelerate meaningful adoption of AI/ML approaches that support model-informed decision-making. PAGE participants are invited to join the SIG, whether to take on leadership roles and shape best practices, contribute to working groups, or simply stay updated on the latest AI/ML methodologies and developments in pharmacometrics across the global community. Through collaborative efforts and shared resources, the SIG aims to support the pharmacometrics community in adopting AI/ML approaches that balance innovation with scientific rigor and patient safety.
References:
[1] FDA & EMA. Guiding Principles of Good AI Practice in Drug Development. U.S. FDA & European Medicines Agency; January 14, 2026
[2] Ponce‐Bobadilla, A. V., Bräm, D., Farnoud, A., Fröhlich, H., Janssen, A., Korsbo, N., … & Mensing, S. (2025). Predictive AI in Clinical Pharmacology: A Call to Action to Develop Robust Benchmarking Practices. CPT: Pharmacometrics & Systems Pharmacology, 15(1), e70155.
Reference: PAGE 34 (2026) Abstr 11999 [www.page-meeting.org/?abstract=11999]
Poster: Methodology – AI/Machine Learning