Google's AI Can Now Detect Breast Cancer Better Than Humans Do

Monday, 15 October 2018 - 11:46AM
Technology
Medical Tech
Artificial Intelligence
Monday, 15 October 2018 - 11:46AM
Google's AI Can Now Detect Breast Cancer Better Than Humans Do
< >
Composite from Pixabay
When it's not teaching AI how to create fake photos that look totally real, Google and its associated companies are taking big strides in adapting artificial intelligence to the medical industry. Previously, Google taught artificial intelligence systems to predict a patient's likely death date, but now they're taking a slightly more proactive approach to treatment: they've taught deep-learning AI how to spot breast cancer.

In a new study, titled "Artificial Intelligence Based Breast Cancer Nodal Metastasis Detection," researchers outline how an AI called LYNA (short for "Lymph Node Assistant") can be trained how to spot breast cancer that has metastasized (spread beyond its original location) and compromised the lymph nodes, which are a system of tubes spread across the body that help filter and fight harmful substances. Being able to identify how far a cancer has spread is the key to "TNM staging," a process that assigns various "stage" classifications (like Stage I, Stage IV, etc.) to reflect how serious a given cancer has become, a patient's survival odds, and what sort of treatment would be best.

Figuring out whether the cancer has spread to the lymph nodes is a key part of the staging process, but it can be a surprisingly tricky thing to judge. According to the official blog post by Google, 1 in 4 metastatic lymph nodes stage assessments (which judge how many lymph nodes have been affected by cancer) would be revised in retrospect, and only 38% of small metastases were spotted at all by pathologists, who were given limited time to examine various slides. This is where LYNA comes in. According to the Google AI blog:

"...LYNA was able to correctly distinguish a slide with metastatic cancer from a slide without cancer 99% of the time. Further, LYNA was able to accurately pinpoint the location of both cancers and other suspicious regions within each slide, some of which were too small to be consistently detected by pathologists. As such, we reasoned that one potential benefit of LYNA could be to highlight these areas of concern for pathologists to review and determine the final diagnosis."

On top of this incredible accuracy, a separate study conducted by Google found that its AI technology can help pathologists cut the amount of time it takes them to review slide samples by half. It's a stunning example of how artificial intelligence can be adapted to almost any problem, and will hopefully keep millions of people from falling victim to undetected cancers.
Science
Science News
Technology
Medical Tech
Artificial Intelligence
No