Digital pathology and AI firm Paige has announed the launch of OmniScreen, a pioneering AI-driven biomarker module capable of evaluating over 505 genes and detecting 1,228 molecular biomarkers from routine H&E-stained digital pathology slides.
Built on Paige’s second-generation foundation model, trained on an unprecedented three million slides, this breakthrough represents a major leap forward in the accuracy, speed, and cost-effectiveness of cancer diagnosis and treatment selection.
Unlike traditional methods that require separate models for each biomarker or cancer type, Paige OmniScreen employs a single AI module capable of predicting a broad spectrum of clinically relevant molecular biomarkers across multiple cancer types. By replicating a targeted biomarker gene panel of 505 genes, it can simultaneously predict 1,228 biomarkers in the 15 most common cancers, including actionable targets such as BRAF, EGFR, KRAS, MET and FGFR3. Additionally, it links phenotypes with genetic patterns, simplifying the diagnosis of conditions defined by specific genetic markers. This capability unlocks deeper, more comprehensive insights into cancer biology and paves the way for new treatment strategies.
This technology accelerates drug development by identifying new potential indications through pan-tumour screening for targeted genetic mutations and enhances patient selection for clinical trials. By pre-screening patients before costly molecular tests, Paige OmniScreen offers substantial cost savings for clinical trials and laboratories, ultimately making personalised therapies more accessible and affordable for a broader range of patients.
“The development of our digital biomarker panel represents a significant advancement in personalised medicine, addressing many of the challenges associated with traditional molecular biomarker tests,” said Razik Yousfi, CTO and CEO of Paige. “Our innovative approach minimises the need for extensive tissue samples, making it particularly useful in cases where biopsy tissue is limited. This first of its kind AI-module offers detailed insights into cancer biology and potential treatment avenues by predicting the activity of canonical signalling pathways, DNA repair mechanisms, and genomic instability measures.”