A paper by FDA authors [Hao et al., 2026] systematically evaluated the contribution of chronic 6-month non-human primates (NHP) toxicity studies to human safety assessment for monoclonal antibodies (mAbs). Three key points stand out:
• Long-term NHP toxicity studies rarely altered regulatory conclusions regarding human risk.
• Long-term studies generally did not identify unexpected off-target toxicities, which is one of the primary reasons these studies are performed in the first place.
• Most findings were predictable from pharmacology, related to immunogenicity, or mechanistically explainable from shorter-duration studies.
This is consistent with a growing body of literature building the same case, including work by Chien et al., 2023, Rana et al., 2024, and Hao et al., 2026. These findings strongly support shifting to a Weight of Evidence (WoE) framework, where the decision to conduct a chronic NHP study is made by design rather than by default. Under a WoE framework, the question becomes: given everything already known about the pharmacology, mechanism of action, platform history, and existing short-term data, does a 6-month NHP study add information that cannot be obtained any other way? For most mAbs, the honest answer is increasingly no. The FDA has released a draft guidance that addresses this directly: A new guidance eliminates routine 6-month primate toxicology studies for monoclonal antibodies. If the field is moving toward WoE-based decision making, the quality of the evidence being weighed becomes critical. This is where model-informed drug development (MIDD) has a natural and underutilized role.
MIDD approaches can contribute directly to WoE assessments in ways that are both mechanistic and human-relevant MIDD approaches could help:
• Predict systemic and tissue exposures
• Integrate target expression in tissues
• Incorporate binding kinetics and target turnover
• Leverage data from in vitro human-cell systems and mechanistic assays
• Quantify translational PK/PD relationships
• Support human-relevant risk prediction earlier in development
The case studies in Mehta et al., 2025 show how PBPK and QSP models have already supported translational decisions that reduced animal study burden without compromising safety assessment. These are not theoretical capabilities. They are being applied now.
What is still missing is published evidence on MIDD as a formal WoE component for NHP study decisions specifically. The mechanistic rationale is clear. The regulatory momentum is building. But the cross-functional collaboration between toxicology, DMPK, translational science, and modeling teams needed to embed MIDD into these decisions consistently is not yet standard practice. That conversation is worth having openly. Are you already incorporating MIDD approaches into WoE strategies in your organization? Do you see a path to reducing chronic NHP reliance through quantitative modeling? The field would benefit from more shared experience on what is working and where the gaps remain.