VivoSim Platform Predicts Gastrointestinal Toxicity (Diarrhea) in an AI Model Trained on Human Intestinal Model Results – Without Animal Testing

SAN DIEGO, April 28, 2026 (GLOBE NEWSWIRE) -- VivoSim Labs, Inc. (NASDAQ:VIVS) (the "Company" or "VivoSim"), a provider of next-generation New Approach Methodologies (NAMs) for preclinical safety, today announced the availability of an AI prediction tool leveraging its NAMkind™ intestinal models to accurately predict the potential of a given drug compound to cause diarrhea in patients. The tool integrates its proprietary NAMkind™ ileum and colon tissues with advanced machine‑learning analytics to identify drug‑induced disruptions to intestinal epithelial integrity and function. Tissue-based assay data is used to train an AI predictive model in a process VivoSim has named VitroSense™.

The model was built using a training set of dozens of compounds. Using high-quality real-world 3d NAM assay results generated from the set of training compounds, the model attained a predictive accuracy of 96% for potential diarrhea. This success represents a significant step forward for the field, demonstrating the value of combining data from VivoSim's complex 3D cell-based assays with machine learning approaches. For a novel compound being investigated for the first time, such as from a client, multiple data endpoints are measured in NAMkind™ intestine assays and these data are fed into the AI prediction model to generate an overall assessment of diarrhea risk. Because so many cancer drugs have dose-limiting toxicities, the Company believes that this breakthrough over reliance on animal models can have a transformational impact on the next generation of cancer treatments.

VitroSense™ – the use of NAMkind™-produced data to train predictive machine learning models

Central to the concept of use of NAMkind™ models in this way is its integration with its proprietary VitroSense™ toolkit that generates trained machine language models using data from its highly relevant human cell NAMkind™ tissues.

"The high biological fidelity of NAMkind™ models allows us to generate data in line with what is seen in humans" said Keith Murphy, VivoSim Executive Chairman. "By using that high-quality data to train AI models, we get much more accurate predictions. You've heard of "garbage in, garbage out", but our models are the opposite: Golden raw data from NAMkind™ tissue assays is fed to the VitroSense™ model and spun into solid gold predictions for our clients' compounds."

NAMkind™ intestine models allow for mimicking oral or IV administration route, and are excellent for antibody drug conjugate testing.

Because of the advanced nature of the NAMkind™ intestinal model, the multilayered structure can be dosed either by mimicking oral administration or intravenous administration by exposing the relevant portion of the tissue – the epithelial lining or the stromal layer.

As recently demonstrated at Society of Toxicology, NAMkind™ intestine models were also tested with antibody-drug conjugates (ADCs) and have the ability to detect differential effects such as antibody activity on epithelium, payload impact on epithelium, and overall ADC impact on epithelium. Permeability endpoints are sensitive to the exact chemical compound, be it ADC, antibody alone, or payload.

"With 96% prediction accuracy, we believe our intestinal model is world class," said Amar Sethi, Chief Scientific Officer of VivoSim. "It has the ability to weed out novel compounds with toxic liabilities from those compounds that would survive rigorous clinical trials."

VivoSim is now working with clients to provide an effective screen for effects on the intestinal epithelium for oncology ADC candidates.

NAMKind™ liver and small intestine toxicology services are now available in US, Europe, and via local distributor engagement across Korea and China, with VivoSim continuing to scale capacity to support expanding global demand and urgent, real-world development needs.

About VivoSim Labs

VivoSim Labs, Inc. ("VivoSim" and the "Company"), is a pharmaceutical and biotechnology services company that is focused on providing testing of drugs and drug candidates in three-dimensional ("3D") human tissue models of liver and intestine. The Company offers partners liver and intestinal toxicology insights using its new approach methodologies ("NAM") models. The Company anticipates accelerated adoption of human tissue models following the U.S. Food and Drug Administration ("FDA") announcement on April 10, 2025 to refine animal testing requirements in favor of these non-animal NAM methods. VivoSim Labs operates from San Diego, CA. Visit www.vivosim.ai.

Forward-Looking Statements

Any statements contained in this press release that do not describe historical facts constitute forward-looking statements as that term is defined in the Private Securities Litigation Reform Act of 1995. Any forward-looking statements contained herein are based on current expectations but are subject to a number of risks and uncertainties. Forward-looking statements include statements regarding NAMKind™, including target turnaround time and its potential to help users de-risk their pipelines, avoid costly downstream failures, reduce rework, prioritize the right assets, move faster, save millions and reduce risk; VivoSim's commercial presence across Asia-Pacific; the evaluation and acceptance of scientifically robust NAM-based evidence; the Company's ability to capture growing demand in the in vitro toxicology testing market; demand for human-relevant toxicology; the market opportunity and market size of gastrointestinal in vitro models and toxicology services; and the Company's scaling capacity to support expanding global demand and development needs. Such forward-looking statements are not guarantees of performance and actual actions or events could differ materially from those contained in such statements. These risks and uncertainties and other factors are identified and described in more detail in the Company's filings with the SEC, including its Annual Report on Form 10-K filed with the SEC on June 5, 2025, as such risk factors are updated in its most recently filed Quarterly Report on Form 10-Q filed with the SEC on February 11, 2026. You should not place undue reliance on these forward-looking statements, which speak only as of the date that they were made. These cautionary statements should be considered with any written or oral forward-looking statements that the Company may issue in the future. Except as required by applicable law, including the securities laws of the United States, the Company does not intend to update any of the forward-looking statements to conform these statements to reflect actual results, later events, or circumstances or to reflect the occurrence of unanticipated events. 

Contact(s):

Investor Relations

info@vivosim.ai

VivoSim Labs, Inc.