FDA Roadmap to Reducing Animal Testing: A Step Toward Smarter Tools for in silico Immunogenicity Risk Assessment

The FDA’s new Roadmap to Reducing Animal Testing in Preclinical Safety Studies lays out a phased plan to move away from animal models and toward approaches that are more predictive of human biology. Central to the roadmap is the advancement of New Approach Methodologies (NAMs), a category that includes in silico modeling and simulation, in vitro systems using human cells or tissues, and integrated platforms that combine both. These methods offer a way to generate meaningful preclinical safety data while minimizing or replacing animal use. 

The document recognizes that animals, especially in the case of biologics like monoclonal antibodies, often fail to accurately model human immune responses. Interspecies differences in immune systems can make it difficult to interpret toxicity findings and can obscure immunogenicity risks that are relevant to patients. In contrast, NAMs are designed to be more physiologically relevant to humans and can be applied earlier in development, when decisions about candidate viability are still being made. 

The FDA is also encouraging the use of integrated testing strategies that combine multiple NAMs to achieve the same objectives as traditional animal studies. For example, pairing organ-on-a-chip systems with physiologically based pharmacokinetic (PBPK) models and AI-driven in silico immunogenicity predictors can provide a more comprehensive and human-relevant view of a candidate’s safety profile, without relying on animal data. These approaches offer the potential for greater predictive accuracy and improved translational relevance.  

As a research organization that has pioneered the use of in silico immunogenicity screening and risk characterization approaches for biologics, EpiVax is committed to integrating AI-driven immunogenicity predictions and finetuning in vitro immunogenicity assays, aligning closely with the FDA’s vision. (Read on to learn how the latest upgrades to our tools make them even more powerful!) 

EpiVax had already submitted its in silico immunogenicity screening methodologies in the context of generic peptide drugs as a Drug Master File (DMF), supporting preparation and submission of ANDA-filings well before the release of the new FDA roadmap. We are thrilled to see the FDA’s continued vision to embrace AI across therapeutic modalities.   

The importance of human-relevant models is nothing new for the EpiVax team. We’ve been modeling human immunogenicity with in silico and in vitro methods since 1998. Over the years, we’ve expanded and refined these models to help drug developers answer critical questions that arise throughout preclinical development. 
 

Our flagship ISPRI™ Toolkit allows drug developers to directly access a suite of in silico tools to screen for immunogenicity risk during drug discovery and preclinical stages. The current version of ISPRI combines machine learning and sequence-based risk assessment to identify and mitigate potential immunogenicity liabilities long before a therapeutic enters the clinic and generates valuable data for quick decision-making and proactive planning for clinical immunogenicity strategy.  

During early-stage lead discovery, ISPRI can be used to help downselect a large panel of hits to a smaller number of leads, using immunogenicity as a key quality attribute. Tech-savvy clients may also leverage the ability to arrange an application programming interface (API) to enable automation of high-throughput screening, saving teams valuable time and resources. In the lead selection phase, developers can use ISPRI to develop a comprehensive data profile demonstrating a thorough immunogenicity risk assessment, which is key to informing in-licensing and funding questions on the immunogenicity topic, as well as positioning programs well for regulatory discussions and questions regarding the immunogenic risk (or lack thereof) for a candidate biologic. EpiVax’s in silico reports are often included in IND-filings, and in silico data generated in ISPRI can also inform development of an Integrated Summary of Immunogenicity for clinical stage programs.  

When researchers need to de-risk a promising biologic demonstrating unacceptable levels of predicted risk, ISPRI makes it easy to identify and modify high-scoring or non-human clusters that may contribute to immunogenic risk. Developers can model point mutations within the ISPRI platform without introducing new epitope content and see the effect of the predicted amino acid changes in real time.  

As regulatory expectations evolve, teams that embrace and adopt predictive, human-relevant in silico tools such as ISPRI early in development will be better equipped to make informed decisions and avoid unnecessary risks. EpiVax is here to support that shift, with deep scientific expertise and tools that fit seamlessly into existing workflows and improve decision-making. 

EpiVax Strengthens ISPRI™ Platform with Next Generation in silico Models for Accurate Characterization of T cell Epitope Content and Clinical ADA Prediction

EpiVax has introduced a series of significant updates to the ISPRI™ platform, its industry-leading suite of tools for the in silico immunogenicity risk assessment of biologics. ISPRI is powered by EpiVax’s proprietary algorithms, including EpiMatrix® for HLA binding prediction and JanusMatrix for assessing T cell epitope cross-conservation. These tools work together to provide a detailed picture of how a biologic candidate may interact with the immune system, supporting more informed decision making across the development pipeline. 

To improve prediction of anti-drug antibody (ADA) responses, EpiVax first built a new version of JanusMatrix (JanusMatrix 2.0) to improve identification of tolerated epitopes using AI techniques. This has enabled more accurate differentiation between foreign and tolerated T cell epitope content, resulting in more effective immunogenicity risk assessments.  

These advancements were key to the development of a new clinical ADA prediction model (ADA model 2.0), trained on a large dataset of therapeutic antibodies. By leveraging AI and ML techniques, the updated model integrates JanusMatrix 2.0 as a core parameter and demonstrates significantly improved performance, increasing predictive accuracy from 62.7% to 87.3%. It also greatly reduces the number of antibodies with underpredicted ADA, offering developers a more reliable way to flag immunogenic candidates and reduce risk earlier in development.  

These updated tools will soon be available within the ISPRI platform. As a key component of the drug development pipeline, ISPRI continues to evolve with industry needs, delivering high-value insights to partners in biotech and pharma worldwide. We recently held a webinar in partnership with The Antibody Society where our newly appointed Chief Technology Officer, Dr. Guilhem Richard, presented these updates in detail.

These latest advancements come at a pivotal time, as regulatory agencies encourage the adoption of new, human-relevant approaches to assessing the safety of therapeutics. With the updated tools in place, ISPRI offers even more accurate assessments not just for antibodies and ADA risk, but for a broad range of biologic modalities. Whether a team is screening leads, preparing an IND or looking to reduce their reliance on in vivo testing, these tools are designed to deliver predictive insights that align with the goals of the FDA’s new roadmap.  

EpiVax Announces New CSO

Thank you for staying connected with EpiVax. We’re proud to be a trusted partner to pharma and biotech teams worldwide, and we’re ready to help developers move their pipelines forward with confidence.  

Questions?