An artificial intelligence proved to be more efficient than human doctors when it comes to correctly identifying brain tumors.
With constant innovations and exciting technological advancements being made every day, it seems that future holds bright things, especially when it comes to medicine. One of those promising technologies is artificial intelligence, which could transform the way diagnose disease- and that’s just one of its possible applications in the health industry.
An international team of scientists succeeded in giving us an early glimpse of what machine-learning algorithms could offer- and even at this stage, the results are more than impressive. They managed to create and train an AI to diagnose over a 100 types of brain tumors with staggering accuracy. In fact, the AI even managed to correctly identify the brain tumor in 12% of cases where real doctors made a misdiagnosis.
For their machine-learning algorithm to have results like these, researchers fed it data from 2,800 cancer patients and enabled the AI to differentiate the tumor types based on a mechanism called DNA methylation. When they gave the AI its first 1,000 tumors to identify after creating a knowledge base, researchers first thought that the 12% difference in results was a mistake on the machine-learning algorithm’s part. As they realized that the AI actually managed to see what escaped our human eyes, they immediately knew that there is an immense potential for this system to change countless lives. Which is why they’ve made it available to the public: without any costs.
Yes, you’ve read that right, the AI that can diagnose types of brain cancer with astonishing precision is free for anyone to use. Adapted to a free online classifier tool, this machine-learning algorithm has been successfully used in 5,000 cases. The goal for the researchers is to “fundamentally transform tumor pathology,” and not only in the field of neurooncology but for all types of cancer.
Photo credit: Mopic/Shutterstock