Japanese researchers use AI to diagnose stomach cancer
AI in diagnostics

Japanese researchers use AI to diagnose stomach cancer

A breakthrough in the use of AI in diagnostics has seen two Japanese national research institutes use artificial intelligence (AI) to identify early-stage stomach cancer.

The Japan Times reports that the AI developed by Riken and the National Cancer Centre took just 0.004 seconds to judge whether an endoscopic image showed early-stage cancer or normal stomach tissue.

The AI was able to match the diagnosis abilities of an experienced doctor, correctly detecting cancer in 80 percent of cancer images, while displaying 95 percent accuracy for healthy tissue images.

The team of researchers employed 100 endoscopic images of early-stage stomach cancer and 100 images of normal stomach tissue to test the AI, following its deep learning training.

Early signs of stomach cancer are extremely difficult to detect, often bearing close resemblance to inflammation. Yet early detection is vital to giving patients the best chance of surviving the disease.

The researchers now aim to make their AI available to doctors to support them in making diagnoses.

Internet of Business says

2012 research revealed that Japan had the third highest rate of stomach cancer in the world (29.9 per 100,000) – around four times higher than the UK. According to Japan’s National Cancer Centre, 45,531 people died of the disease in 2016 alone.

AI therefore represents a key avenue for tackling stomach cancer in Japan by enabling earlier, more accurate, and faster diagnoses.

AI is already gaining traction in diagnosing other cancers, thanks to companies such as Lunit, which has developed an AI for detecting lung and breast cancer, and Ultromics, which is using artificial intelligence to help doctors diagnose heart disease and lung cancer.

Many people believe that AI could also be the saviour of an NHS beset by budget problems. Sir John Bell, chairman of the Office for Strategic Coordination of Health Research (OSCHR), told BBC News:

There is about £2.2 billion spent on pathology services in the NHS. You may be able to reduce that by 50 percent. AI may be the thing that saves the NHS.

This sentiment was reflected in our own deep dive into how AI could transform the NHS earlier this year, in which we examined the NHS’s current forays into AI and what the future holds in the field. The technology not only has the potential to save huge sums of money, but also to assist over-stretched doctors, and ultimately to improve the quality of care.

Perhaps the most famous proponent of AI in health is Google offshoot DeepMind Health, which has been collaborating with the NHS to apply its AI to eye disease diagnosis.

However, the company has also fallen foul of privacy regulations: an ongoing risk when applying AI to large volumes of patient data. A panel of independent reviewers – commissioned by DeepMind Health itself – emphasised the problem last month, when they warned that the company could exert “excessive monopoly power” in the healthcare and medical data spaces.

But while questions remain over how technology companies will be able to access and manage patient data without falling foul of data protection rules, there’s no doubt that AI has a big future in healthcare, with diagnostics leading the way.