Page and Microsoft announced Virchow2 and Virchow2G, enhanced versions of their groundbreaking AI model for cancer pathology, as the next major advancement in clinical AI for cancer diagnosis and treatment.
The Virchow2 and Virchow2G models are based on a massive dataset accumulated by Paige, which collected over 3 million pathology slides from over 800 laboratories across 45 countries and trained the models on them. Not surprisingly, such a large amount of data is extremely useful. The data comes from over 225,000 patients, all of which have been anonymized to create a rich and representative dataset that encompasses all genders, races, ethnicities, and regions around the world.
What makes these models truly remarkable is their scope: they cover over 40 tissue types and a variety of stains, making them applicable to a wide range of cancer diagnostics. With 1.8 billion parameters, Virchow2G is the largest pathology model ever created, setting a new standard for AI training, scale, and performance.
Dr. Thomas Fuchs, founder and chief scientist at Paige, commented: “We are only just beginning to explore what these foundational models can achieve in fundamentally changing our understanding of cancer through computational pathology.” Dr. Fuchs believes these models will greatly improve the future for pathologists, and agrees that the technology is shaping up to be an important step in the advancement of diagnostics, targeted medicines and customized patient care.
Similarly, Raziq Yousfi, senior vice president of technology at Page, said these models are not only enabling precision medicine, but also improving the accuracy and efficiency of cancer diagnostics, pushing the boundaries of what’s possible in pathology and patient care.
So how does this relate to cancer diagnosis today? Page has developed a clinical AI application that helps pathologists recognize cancer in more than 40 types of tissue. The tool allows them to identify potentially dangerous areas faster and more accurately. In other words, the tool makes the diagnostic process more efficient and less prone to errors, even for rare cancers.
Beyond diagnostics, Page has developed AI modules that are useful to life science and pharmaceutical companies: These tools can aid in therapeutic targeting, biomarker identification, and clinical trial design, potentially improving the success of clinical trials and speeding up the development of new treatments.
The good news for researchers is that Virchow2 is available on Hugging Face for non-profit research, and the full suite of AI modules is now available for commercial use. This accessibility has the potential to accelerate advances in cancer research and treatment across the scientific community.
Taken together, the recently introduced AI models represent a major breakthrough in the fight against cancer. Page and Microsoft have chosen the right path by combining the power of data with cutting-edge AI technology. Together, the two companies have created new opportunities for more accurate cancer predictions, paving the way for customized solutions and innovative research in oncology.
(Photo courtesy of the National Cancer Institute)
Read also: The hidden climate costs of AI: Tech giants struggle to go green
Want to learn more about AI and Big Data from industry leaders? Check out the AI & Big Data Expo in Amsterdam, California, and London. This comprehensive event will be held alongside other major events such as Intelligent Automation Conference, BlockX, Digital Transformation Week, and Cyber Security & Cloud Expo.
Find out about upcoming enterprise technology events and webinars hosted by TechForge here.
