Technology
This AI-driven technology predicts omics signatures directly from histological biopsy images, transforming standard pathology into a multi-omics-powered diagnostic tool. The method integrates machine learning to map molecular data onto histology, enabling precise prognosis and personalized treatment decisions.
Tested on prostate cancer data from 401 patients, the approach successfully integrated histopathology, transcriptome profiling, CNV, SNV, and clinical metadata to predict remission outcomes. The combination of genetic, RNA expression, and histology data significantly improved prognosis. Additionally, in Eosinophilic Esophagitis (EoE), the method analyzed over 1,000 histology images to predict omics states, demonstrating its potential for assessing disease severity.
By combining RNA sequencing, DNA methylation, and protein expression with digital pathology, this approach enhances cancer diagnostics, particularly for breast and prostate cancer, and improves biopsy-based prognostics.
Advantages
- Enhances Standard Biopsy Analysis – Predicts molecular signatures without additional sequencing.
- Personalized Medicine – Identifies patients who benefit most from omics testing.
- Validated in Cancer & GI Diseases – Proven efficacy in prostate cancer and EoE.
Applications & Opportunities
- Cancer Diagnostics – Applied to 14 cancer types for prognosis improvement.
- Gastrointestinal Diseases – Enhances histology-based severity prediction for immune-related conditions like EoE.
- Facilitate explainability of drug response trials along the drug development trajectory.
Business Development Contacts