Xellar Biosystems and Pfizer Unveil Next-Gen AI-Powered Image Decision Engine, Accelerating Digital Pathology into the Sub-Second Era with 100x Efficiency Gains

[Boston, June 2025] Recently, the global next-generation AI-driven organ-on-a-chip platform innovation companyXellar BiosystemsCompared with international pharmaceutical giantsPfizerJointly Release a Major Achievement:AI Deep Learning Pathology Image Automatic Recognition System。

This system has successfully achieved precise recognition and automatic quantification of pancreatic tissue structures. It marks the transition of non-clinical toxicity assessment from experience-dependent to data-driven, laying an important foundation for the large-scale application of AI in human-derived models.This platform can be widely applied in multiple key areas such as drug toxicity modeling, pathological mechanism analysis, and early response interpretation of target organs, providing a higher-resolution "imaging decision engine" for the non-clinical stage.

The AI pathology recognition system developed in this researchCompressing the traditional process, which requires hours of manual annotation, to be completed within 30 seconds, achieving a "hundredfold increase in efficiency" leap forward.More importantly, the system not only achieves consistency comparable to or even better than human performance, but also successfully transforms visual information, previously only used for "reference," into structured image data assets that are "quantifiable, archivable, and computable."This provides a new data foundation for critical applications such as toxicity mechanism research, dose-dependent modeling, and target organ response monitoring, and also establishes a technical paradigm for the implementation of AI in Non-Animal Methods (NAMs) systems.

This work not only focuses on the precise modeling of pancreatic structures but also demonstrates the powerful scalability of AI in multi-organ, multi-modal pathological analysis.The output format and algorithm design of this system fully comply with the latest requirements of international regulatory agencies such as the FDA and OECD for "verifiability, repeatability, and transparency," and are expected to become a core component in the future validation of NAMs and AI-assisted toxicology pathways.

Dr. Xie Xin, Founder and CEO of Xellar BiosystemsSaid: "Image data is one of the most intuitive yet complex sources of information in toxicology research. We hope that by empowering with AI, these 'visual pieces of information' can truly become computable, summarizable, and actionable knowledge resources. The work with the Pfizer team not only represents a scientific breakthrough but also holds significance for global regulation and application."

Dr. Zhiyong Xie, Vice President of AI at Xellar BiosystemsHe added: "The integration of AI and pathology allows us to 'quantify' the toxic mechanisms of drugs from images, rather than just observing them. What we are driving is not only technological innovation but also a standardization revolution in the R&D process."

The research findings have recently been published in an international authoritative journal.《Toxicologic Pathology》, becoming a key driver of "AI-Driven Digital Toxicology"A key link in the strategy."

Industry and Technology Significance

• From "Image Viewing" to "Image Computing": A New Paradigm for AI Pathology

This work has achieved deep learning recognition and automatic quantification of the complex structure of the pancreas for the first time, transforming traditional subjective pathological judgment into a standardized, data-driven computational process. This "image quantification" capability can be widely adapted for capturing early toxicity signals and dose-response modeling in drug development.

• Promote the Regulatory Transformation of Non-Animal Methods (NAMs)

As international regulatory agencies such as the FDA, EMA of the European Union, and OECD continue to promote the construction of non-animal evaluation systems, AI pathology systems serve as a key fulcrum in the NAMs data chain, providing forward-looking validation samples for global digital toxicology regulatory compliance.

• Scalable Platform for Multiple Organs

In the future, this platform can be quickly expanded to multiple high-risk target organs such as the liver, kidney, and central nervous system, providing "structural-level" AI insights for the safety evaluation throughout the drug lifecycle, while supporting industry trends like the 3R policy and green research and development.

In the future, Xellar Biosystems will continue to expand the application of AI in multiple directions such as drug toxicity, efficacy modeling, and patient stratification, deepen strategic collaboration with international pharmaceutical companies, and accelerate the construction of a new generation of intelligent new drug R&D platform centered on human-derived data.

Under the convergence of AI, regulatory reform, and scientific validation, the joint research achievement by Xellar Biosystems and Pfizer is not only a scientific breakthrough but also a significant milestone in advancing the global drug safety evaluation system from "animal models" to "data-driven human-based systems."

Recommended Reading