New research collaboration applies foundation models to building AI solutions for accelerated preclinical safety assessment.

BOSTON: Tremont AI, an AI company specialized in preclinical drug safety, today announced a research collaboration with AstraZeneca to develop and evaluate AI solutions for accelerated safety screening in preclinical toxicologic pathology studies.

Assessing the safety of a new drug is one of the most demanding steps before it can reach patients. Toxicologic pathologists review vast numbers of images of tissue exposed to the candidate compound, looking for treatment-related changes that must be told apart from normal biological variation. The judgments are critical to determining the compound's safety profile and the dose at which it can be given safely. Yet today the work depends on expert interpretation that is semi-quantitative and subject to variability. Foundation models can turn this interpretation into a measurable and reproducible process. Tremont AI is exploring how foundation models can strengthen preclinical drug safety assessment, freeing pathologists from repetitive review to focus on interpretation.

"Foundation models can change how preclinical safety data is read, making it faster to triage and more consistent to interpret," said Luca Weishaupt, Chief Executive Officer and co-founder of Tremont AI. "Working with AstraZeneca is enabling us to enhance and evaluate these models against the realities of drug safety assessment, where findings can be rare and difficult to discern from background findings. These are conditions under which general-purpose models tend to break down, and they are exactly what toxicology-specific foundation models are built to handle."

About Tremont AI

Tremont AI is a Boston-based biomedical AI company building foundation models and agentic AI for preclinical toxicologic pathology. Spun out of the Mahmood Lab at Mass General Brigham, the company develops multimodal foundation models and integrated solutions that help pharmaceutical companies and contract research organizations make preclinical drug safety assessment faster, more consistent, and more computable. For more information, visit tremont.ai.

Forward-Looking Statements

This press release may contain forward-looking statements that reflect current assumptions and forecasts made by Tremont AI management regarding future events. Actual results may differ materially. Tremont AI does not assume any obligation to update these statements in response to future events.

Fonte: Business Wire