Microsoft Enlists AI, Cloud in Battle Against Cancer
Microsoft researchers are using artificial intelligence technologies and the Azure cloud computing platform to advance cancer research.The American Cancer Society estimates that there will be nearly 1.7 million new cancer cases diagnosed in the United States and more than 595,000 cancer deaths this year. Microsoft hopes to put a big dent in those numbers by using its research in artificial intelligence (AI) and its massive, globe-spanning Azure cloud to "solve" cancer. The company's researchers are exploring ways to help oncologists find the most effective cancer treatments and tailor them to individual patients using the latest innovations in machine learning and natural language processing. In a Microsoft Stories post, senior writer Allison Linn detailed some of the technologies the company is bringing to bear against cancer. Microsoft is exploring multiple approaches, including a cloud-based tool called the Bio Model Analyzer, which creates a model of healthy cells and compares it against cells that grow cancerous, evaluating the interactions in the genes and proteins, offering researchers a window into how cancer spreads. The technology could help detect cancer earlier and may help doctors design effective treatments, cope with rare cancers and determine when and if a cancer will become resistant to a given treatment. Another cloud-based system called Literome trawls through millions of research papers, enabling oncologists to spend more time easing their patients' suffering and less time combing through thick volumes of medical research. "To build Literome, [Microsoft researcher Hoifung] Poon and his colleagues used machine learning to develop natural language processing tools that require only a small amount of available knowledge to create a sophisticated model for finding those different descriptions of similar knowledge," wrote Linn.
Microsoft is also exploring how to analyze CT (computerized tomography) scans—up to 2,000 images in some cases—to spot tumors. One research project involves using computer vision and machine learning to perform a pixel-by-pixel evaluation of 3D CT scans to determine how much a tumor has shrunk, grown or changes shape between scans.