The tool is called MUSK, short for multimodal transformer with unified mask modeling, and it outperformed standard prediction models across a variety of cancer types, according to a Jan. 8 news release from Stanford Medicine.
Here are five things to know about the tool:
- The model was trained on 50 million medical images of standard pathology slides and more than 1 billion pathology-related texts and associated follow-up data from The Cancer Genome Atlas.
- Sixteen major types of cancer were represented in the data, including breast, bladder, colorectal, head and neck, kidney, lung and pancreatic.
- Across all cancer types, the AI model accurately predicted patient survival 75% of the time, compared to accurate predictions 64% of the time for standard models.
- For non-small cell lung cancer, the AI model accurately identified patients who benefited from immunotherapy treatment 77% of the time, compared to accurate predictions 61% of the time for standard models.
- For melanoma, the AI model accurately identified patients who were most likely to relapse within five years of treatment 83% of the time, compared to accurate predictions 71% of the time for standard models.
Access the full study here.
At the Becker's Perioperative Summit, taking place September 14–15 in Chicago, perioperative leaders and healthcare executives will focus on improving operating room efficiency, enhancing patient safety, optimizing staffing and driving innovation across surgical services. Apply for complimentary registration now.
