Product Launches
Sophont Launches Holon-S1 to Predict Tumor Biology From Routine Pathology Slides
By News Desk ·
Key takeaways
- Sophont announced Holon-S1, a pan-cancer AI model that predicts spatial gene and pathway activity from routine H&E pathology slides.
- Holon-S1 outperformed two other state-of-the-art models released this year in spatial gene activity prediction and robustness.
- Sophont has raised $9.22 million to date and was founded in February 2025 by Tanishq Mathew Abraham and Paul Scotti.
Sophont announced Holon-S1, a pan-cancer AI model that predicts spatial gene and biological pathway activity from routine hematoxylin and eosin pathology slides, on October 20, 2026. The Sacramento-based medical AI startup said the model is available for research use only, according to PR Newswire.
How Holon-S1 Predicts Tumor Biology
Holon-S1 predicts the spatial activity of genes and biological pathways from routine pathology slides, which Sophont said aims to broaden access to insights that typically require costly and time-consuming molecular testing. The company is inviting biopharma and life sciences companies and research institutions to collaborate on biomarker discovery, therapeutic target research, and analysis of existing and archived patient cohorts.
In a retrospective osteosarcoma patient case study using publicly shared slides, Holon-S1 recovered several insights originally obtained through months of extensive molecular testing, identifying a fibroblast-associated program including FAP, the gene targeted by a therapy the patient responded to. Comparing pretreatment and post-treatment slides, the model quantified increased immune infiltration and decreased cancer proliferation. The model had not been trained on osteosarcoma.
In triple-negative breast cancer, a biomarker defined by the spatial organization of predicted EREG activity was associated with response to chemotherapy before surgery in discovery and separate held-out cohorts. Sophont evaluated Holon-S1 across 16 organ types, more than 15 scanners, and formalin-fixed, paraffin-embedded and fresh-frozen tissue. In a controlled comparison, it outperformed two other state-of-the-art models released this year in spatial gene activity prediction and robustness.
"Most AI pathology solutions infer patient risk profiles without showing clinicians and researchers what biology resulted in the prediction," said Tanishq Mathew Abraham, Ph.D., CEO and Co-Founder of Sophont. "Holon-S1 was designed to uncover the biology behind the answer, showing which genes are active, where in the tumor they are active, and how that is related to whether a patient is likely to respond to treatment."
Sophont's Prior Open Model Releases
Sophont has raised $9.22 million to date, according to akta.pro company records, including a $9.2 million seed round in September 2025 led by Kindred Ventures. The company released OpenMidnight, an open-source pathology foundation model trained for under $1,600, in November 2025, and CortexMAE, an open-source functional MRI foundation model, in May 2026. It also released Medmarks, an open medical LLM benchmark suite, in May 2026.
Other Health AI Launches This Month
Google DeepMind and Google Research released health AI tools including MedGemma, which improved chest X-ray classification accuracy by 15.5-18.1 percentage points, Crypto Briefing reported on October 16, 2026. Neurobit debuted its NeuroSwift Pro AI cerebrovascular assessment platform in Rotterdam on October 13, 2026, with validation across 11 Dutch hospitals, EIN Presswire reported. Spectral AI showcased its FDA-cleared DeepView burn assessment system at the 2026 ACEP Scientific Assembly in Chicago, Stock Titan reported on October 13, 2026.
Further Holon-S1 Studies Planned
Sophont said it is designing further retrospective and prospective studies to evaluate potential diagnostic use of Holon-S1. Details of the model appear in Holon-S1's technical report.