AI News Feed
Microsoft Open-Sources GigaTIME for Tumor Mapping
11 Dec 2025- Microsoft open-sourced GigaTIME, an AI trained on 40M cell samples that converts routine tissue slides into detailed tumor maps and population-scale immune insights, enabling fast, low-cost biomarker analysis.
11 Dec 2025
Microsoft has open-sourced GigaTIME, an AI model that can turn routine, inexpensive tissue slides into rich tumor maps and population-scale insights. Trained on 40 million cell samples provided by Providence Health, GigaTIME learns to match basic slide images with more advanced immune-system scans. In validation work, researchers ran the model across data from more than 14,000 cancer patients, producing a “virtual library” of roughly 300,000 detailed tumor images spanning 24 cancer types. That virtual population allowed the team to surface over 1,200 patterns linking immune activity to clinical factors such as cancer stage and patient survival. The newsletter highlights that the same analysis previously required costly lab assays and days of processing, whereas GigaTIME can generate insights in seconds from a ~$10 tissue slide.
Why it matters: by pulling meaningful, clinically relevant signals from routine pathology data, GigaTIME represents a step toward making population-scale tumor analysis far cheaper and faster. Open-sourcing the model could lower barriers for researchers and clinicians to run large-scale studies, accelerate biomarker discovery, and help inform treatment decisions without the time and cost of specialized lab workflows.
Source
Microsoft has open-sourced GigaTIME, an AI model that can turn routine, inexpensive tissue slides into rich tumor maps and population-scale insights. Trained on 40 million cell samples provided by Providence Health, GigaTIME learns to match basic slide images with more advanced immune-system scans. In validation work, researchers ran the model across data from more than 14,000 cancer patients, producing a “virtual library” of roughly 300,000 detailed tumor images spanning 24 cancer types. That virtual population allowed the team to surface over 1,200 patterns linking immune activity to clinical factors such as cancer stage and patient survival. The newsletter highlights that the same analysis previously required costly lab assays and days of processing, whereas GigaTIME can generate insights in seconds from a ~$10 tissue slide.
Why it matters: by pulling meaningful, clinically relevant signals from routine pathology data, GigaTIME represents a step toward making population-scale tumor analysis far cheaper and faster. Open-sourcing the model could lower barriers for researchers and clinicians to run large-scale studies, accelerate biomarker discovery, and help inform treatment decisions without the time and cost of specialized lab workflows.
Source
The method
The prompts
Helping entrepreneurs and business owners use AI tools for growth and productivity.
Wawa Llama Ltd GB447340593