In breakthrough, Japanese researchers use AI to identify early stage stomach cancer with high accuracy

Two Japanese national research institutes have succeeded in using artificial intelligence to identify early stage stomach cancer with a high accuracy rate.

The breakthrough may help extend the lives of patients in Japan, where stomach cancer is one of the leading causes of death. According to the National Cancer Center, 45,531 people died of stomach cancer in 2016.

According to Riken and the National Cancer Center, it took AI only 0.004 seconds to judge whether an endoscopic image showed early stage cancer or normal stomach tissue. AI correctly detected cancer in 80 percent of cancer images, while the accuracy rate was 95 percent for normal tissue.

The accuracy rates were as high as those of veteran doctors, the institutes said Saturday, adding that they will aim to put AI into practical use as a device to support doctors in making diagnoses.

Stomach cancer causes few symptoms and is often found only after it reaches an advanced stage. At an early stage, even specialists have a difficult time distinguishing the cancer from inflammation.

For the AI project, a team of researchers prepared 100 endoscopic images of early stage stomach cancer and 100 images of normal stomach tissue to test AI capabilities in a method known as deep learning.

The results of a large-scale study released in January 2016 showed that those diagnosed with cancer stand a 58.2 percent chance of surviving another 10 years, according to the National Cancer Center. The survival rate for five years is 63.1 percent.

By degree of progression, however, the five-year rate for all types of cancers found at stage one was found to be 90.1 percent, with the 10-year rate standing at 86.3 percent.

In stage four cases, where cancer has spread to other tissues or organs, the five-year survival rate is only 17.4 percent and the 10-year rate a mere 12.2 percent.

AI is demonstrating its medical worth in other nations as well.

A computer running AI software defeated two teams of doctors in recognizing maladies from magnetic resonance images during the CHAIN Cup in Beijing in June — a contest billed as a world first.

And Britain’s National Health Service is working with the AI company Babylon, which says its medical chatbot already outperforms medical trainees on sample questions for the national doctors’ exams.

The chatbot is a key feature of Babylon’s “GP at Hand” app, which has over 50,000 users in the U.K. The firm also has over 2 million members in its health care service in Rwanda. It is collaborating with the tech giants Samsung and Tencent to expand its app offerings and plans further rollouts around the world.


Leave a Reply

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

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

Facebook photo

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

Connecting to %s

Up ↑

%d bloggers like this: