Microsoft is partnering with the digital pathology company Paige to build the world’s largest image-based artificial intelligence (AI) model to help detect cancer, the companies announced.
The AI model will be used for digital pathology and oncology, configured with billions of parameters to provide a computer vision AI that is orders of magnitude larger than any similar model existing today.
Dr. Thomas Fuchs, Paige’s founder and chief scientist, told FOX News Digital that the amount of data used in the model is “orders of magnitude” larger than anything made public by Google or Facebook.
“It’s so much larger than anything that has been published in that area ever,” he said.
That scale is essential for patients. While already established models include up to 100,000 slides, Fuchs said a model that can accurately assist pathologists and oncologists needs digitized histology slides in the millions.
Paige has helped digitize millions of slides to identify cancers for seven years. Such a task required billions of images and ever-growing AI models, but with such a sizable compute cast, Paige needed more power to work to that scale.
“That’s where Microsoft comes in,” Fuchs said. “We work with Microsoft on several fronts. So, with Nuance, for example, with Azure and with Microsoft Research and their resources and their compute, we can now build models from thousands of GPUs [graphics processing units] used at an unprecedented scale.”
The next development phase will incorporate up to 4 million digitized microscopy slides from a petabyte-scale clinical data archive. With Microsoft’s massive established supercomputer infrastructure, Paige will be able to train the model to deploy across laboratories and hospitals using Azure.
Fuchs considers the AI model “ChatGPT for the microscopic world.”
“It’s the foundation model for the microscopic world,” he added. “The data we use a page to build this foundation model with Microsoft is 10 times more than all of Netflix. All the shows you watch, all the movies, every pixel in all these — there’s hundreds of shows, thousands of hours of content, but is just 10% in terms of data of the images we put through these models. So, this will not only be the largest oncology model — it will be the largest-ever publicly announced computer vision model.”
The benefits of the model will help better understand the morphology of cancer of normal tissue to a meticulous detail, which can then be used as a foundation for rare cancer models, mutation prediction and response prediction. The model is anticipated to reduce misdiagnosis in medicine significantly.
Misdiagnoses in the United States lead to hundreds of thousands of deaths and major disabilities each year, according to a recent report from Johns Hopkins School of Medicine.
Each year, an estimated 795,000 Americans become permanently disabled or die due to a misdiagnosis, the study found.
Eventually, Fuchs hopes the model can be tuned for all rare cancers or build detection systems across it. A prototype already exists that might soon be able to serve any cancer patient, not only a fraction, such as breast or lung cancer.
Paige made waves in the healthcare industry after developing the first Large Foundation Model. The model uses more than a billion images from half-a-million pathology slides comprising multiple types of cancers.
They are the first company to receive FDA approval for a clinical AI application in digital pathology for the model. The same Paige AI technology assists pharmaceutical companies in evaluating effective treatment plans for patients.
“Paige has been at the forefront of innovation since its inception, and by combining Microsoft’s expertise and enormous compute power with Paige’s deep expertise in AI, technology and digital pathology, we strongly believe we will significantly advance the state-of-the-art in cancer imaging,” said Razik Yousfi, SVP of Engineering at Paige. “Through the development of this model, we will help improve the lives of the millions of people who are affected by cancer every day.”
For more Culture, Media, Education, Opinion, and channel coverage, visit foxnews.com/media.