by John R. Fischer
, Senior Reporter | September 13, 2023
Using Microsoft’s supercomputer and cloud-based technologies, digital pathology and clinical AI developer Paige has set to work building the world’s largest image-based, generative AI models for developing next-generation clinical applications and computational biomarkers in digital pathology and oncology.
The company has already created the first Large Foundational Model from one billion images it collected from half a million pathology slides for multiple types of cancer. It will now add up to four million digitized microscopy slides, also for multiple cancer types, using its petabyte-scale archive of clinical data.
Paige says that its AI model will be configured with billions of parameters, making it orders of magnitude larger than any other image-based AI model, and help in identifying the subtle complexities of cancer.
“It opens a window into the microscopic world with extraordinary fidelity, allowing for not only much higher accuracy but completely novel capabilities,” said Thomas Fuchs, founder and chief scientist of Paige, in a statement.
Once built, the company will train the model using Microsoft’s advanced supercomputing infrastructure, and deploy it to hospitals and labs worldwide using Microsoft’s cloud infrastructure platform, Azure
The two companies announced
their intention back in January to create these large AI models by combining Azure with the Paige Platform, a comprehensive solution that uses data management, cloud storage options, and FullFocus, an FDA-cleared whole-slide image viewer, to power laboratory digital pathology workflows.
Paige is currently the only company to have received FDA approval for an AI algorithm in pathology. Through this agreement, Paige is a Microsoft Cloud for Healthcare partner and is working with the company to scale its technology’s global reach and accelerate AI adoption in digital pathology.
It also is working with Nuance, a Microsoft company, to integrate its solutions with Nuance’s Precision Imaging Network, which combines Nuance’s workflows and reporting and image‑sharing infrastructure with the scale, security, and strength of Azure to facilitate seamless hosting, integration, and management of partner AI services on a single platform.