From the Signal blog: GigaTIME AI tool advances cancer research and could unlock access to better treatment

Every minute, four Americans are told they have cancer and start a journey into one of the most complex diseases in medicine. Now, a new AI tool developed by Microsoft Research and covered today in a new Cell paper allows researchers to study the "neighborhood" around tumors at a new scale, which is key to predicting how cancers behave and which treatments work best.

Relying on models that can process multiple types of data, GigaTIME could help change how researchers study cancer. Research often relies on expensive time-consuming lab tests to learn how tumors develop in the body. GigaTIME turns inexpensive everyday pathology slides into complex and detailed digital maps that show how immune cells interact with cancerous tumors through protein activation, enabling researchers to uncover previously invisible patterns.

In this way, GigaTIME opens opportunities to study tumor microenvironments at a scale never seen before. Traditionally, generating these types of digital maps would take days and thousands of dollars for a single sample. By contrast, GigaTIME can simulate such analysis over dozens of proteins in seconds through computation, enabling researchers to study tens of thousands of scenarios at once. Eventually, the tool could help identify which patients might benefit from specific treatments and improve odds for patients. It could also shed light on why some patients fail to respond and how to counter tumor resistance.

GigaTIME is publicly available as an open-source research tool on Microsoft Foundry Labs and Hugging Face. Developed in collaboration with Providence and the University of Washington, it leverages data approved for research use from a diverse cohort of more than 14,000 patients across 51 hospitals and over 1,000 clinics.

Read more in our MSR blog and the paper in Cell.