- Joined
- Mar 3, 2018
- Messages
- 1,713
Machine learning technology is making massive inroads into the medical industry. AI systems are already beating human radiologists at identifying tumors in digital scans, but an article published in Science Translational Medicine claims machine learning algorithms can speed up tumor analysis in labs. Correctly identifying a tumor is critical to properly treating patients, but the difficult process typically requires teams of human reviewers in a centralized lab to ensure reliable results. The software, dubbed "Cerebro", is said to automate the process by generating confidence scores for various mutations. Samuel Angiuoli, co author of the report, says he wants to develop a kit that can "run anywhere in the world" without expensive and remote labs.
Angiuoli and his team trained Cerebro using millions of real-world and in silico mutations. They then pitted Cerebro head-to-head against several existing cancer mutation identification methods and found that the machine learning technique was more accurate in almost every circumstance. "The improvements [to mutation identification software] matter and have clinical implications," says Angiuoli. That's particularly true as more DNA-specific cancer therapies continue to become available on the market, he says.
Angiuoli and his team trained Cerebro using millions of real-world and in silico mutations. They then pitted Cerebro head-to-head against several existing cancer mutation identification methods and found that the machine learning technique was more accurate in almost every circumstance. "The improvements [to mutation identification software] matter and have clinical implications," says Angiuoli. That's particularly true as more DNA-specific cancer therapies continue to become available on the market, he says.