New AI Can Detect a Deadly Cancer Early With 86% Accuracy

Scientists have been attempting to gauge just how much artificial intelligence can be used to supplement–and, in some cases, perhaps even supplant—doctors, radiologists, lab technicians, and other healthcare workers. Now, new research from Japanese scientists suggests AI could be used to detect colorectal cancer in its earliest stages before tumours become malignant and deadly cancer becomes much harder to treat. In fact, the tech was able to detect cancer with 86% accuracy.

Study lead Dr Yuichi Mori of Showa University presented the new data at the United European Gastroenterology conference in Barcelona. Mori and his team collected tens of thousands of high-resolution images of pre-cancerous and cancerous cells in order to kick the machine learning process into gear; their AI algorithm was then able to discern cancers from highly magnified pictures of colorectal polyps within just a second.

The 86% accuracy figure, if it holds steady in future studies, is impressive. Colorectal cancer can become extremely deadly once it’s malignant because it can easily spread to the lymph nodes or bloodstream. However, with early detection, the disease can be treated and prevented from worsening.

AI has also been tested in reading X-rays and brain scans and even narrowing down the list of possible genetic diseases a patient suffers from simply by analyzing his or her face. While the technology may not outright replace human doctors and technicians any time soon, it could serve as a useful tool to help physicians speed up diagnostics.

– SY MUKHERJEE

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s