AI brings changes in medical practice#Oriental Outlook#-Sino-US


Path: Sino-US >> Cover Story>> 2018 >>
AI brings changes in medical practice

In a latest development related to artificial intelligence in the medical sector, Chinese researchers and doctors have developed an AI model with medical imaging to help determine whether patients with severe brain damage can regain consciousness.

Companies have invested heavily on adopting the AI technology in the medical sector and especially in medical imaging, a field where AI can easily assist doctors in analyzing and diagnosing patients’ diseases.

An estimate from industry information provider qianzhan.com shows that China’s medical AI industry scale will reach 20 billion yuan ($2.93 billion) in 2018, a 53.8 percent year-on-year increase.

But it was only after the central government issued a new document on medical equipment in August that the AI medical equipment could gain license to be applied in medical institutions.

In the 35th issue of 2018, the Oriental Outlook magazine under the Xinhua News Agency ran a cover story on the medical AI industry in China, discussing AI advantages in medical sector and barriers to its application.

Below is an excerpt of the article.

Ten minutes, that is the time IBM Watson, an AI medical system, takes to diagnose a rare cancer and suggest the right treatment. Hard to believe!

In recent years, along with the maturing of key AI technologies such as deep learning, image recognition, speech recognition and algorithm capability, many companies have made efforts and invested a lot in the medical AI industry.

A report by McKinsey & Co shows that companies invested about $26 billion to $39 billion in the AI industry in 2016, and it is estimated that the AI application market will be $127 billion by 2025, one-fifth of which will come from the medical sector.

A lot of countries have stepped up AI development in medical sector. China has also joined in. Some Chinese companies have developed products in medical imaging diagnosis, pathological examination and intelligent inquiry. Internet giants Alibaba, Tencent and Baidu have established platforms for medical AI resources.

Since the second quarter of 2015, China saw an explosive growth of investment in the medical AI industry. By August 2017, the total investment in the industry accumulated to over 18 billion yuan, according to information portal iyiou.com.

With the huge capital support, the medical AI technologies have improved a lot in China and there are many areas in which they can be applied.

For example, AI assistant systems would complete primary diagnosis in 30 seconds in general, and the accuracy rate reached 90 percent in pulmonary nodules detection, exceeding the average rate of doctors.

While the medical AI industry is growing at a high speed, there are many barriers it needs to cross.

The commercialization of AI products needs to cross institutional barriers, exploring a more suitable safety review mode; making efforts to let hospitals and doctors accept AI products; making AI products more applicable, simpler and easier to operate, and better match the requirements of hospitals.

It must be admitted that medical AI is still far away from highly intelligent imaging, and even a long way from changing the traditional medical practice.

A more realistic expectation would be that as an assistant product, the medical AI could improve the accuracy in diagnosis and the efficiency in the disease control and health management.

Efficiency

At present, AI technologies have been adopted in at least eight kinds of fields including cytopathology, medical imaging, hospital management and health management, Shan Yang, an analyst from iyiou.com, said.

Zhang Jinglei, vice president of medical AI provider Airdoc, said that from the technology and application perspective, medical imaging has been the most ideal AI application field. His company is focusing on this field.

An AI medical imaging product has found more than 600 patients with fundus oculi diseases, accounting for one-third of total such patients found in half year at the No 2 Community Health Center of Panjiayuan in Beijing’s Chaoyang district.

Wang Zhenchang, vice president of the Beijing Friendship Hospital Affiliated to Capital Medical University, said that AI technologies have brought more and more changes in medical imaging.

Wang said that currently 80 percent of the diagnosis and treatment decisions are based on medical images, such as electrocardiogram and electroencephalogram.

China has only 170,000 registered imaging doctors, with an annual growth of 2 percent, but the imaging data volume is rising by 63 percent annually, according to statistics.

Imaging doctors always work till 9 pm every day at every hospital, said Wang, who has worked on imaging diagnosis since 1986. Eye stress might lead to erroneous diagnosis, he said.

AI systems help locate the targeted areas and make primary diagnosis, which cut 30 percent of time for doctors, preventing mistakes when doctors are suffering from fatigue, said Chen Kuan, CEO of Beijing Infervision, a company focusing on AI applications in medical imaging.

In the pathological examination, since the cervical cancer is the first type of cancer to be discovered at an early stage and cured, many companies are involved in developing AI systems to identify the patients.

About 350 million Chinese women need the examination, but until June 2017, China only had more than 10,000 registered doctors for pathological examination.

According to statistics, only about 5 percent of people undertaking tests might be diagnosed with cervical cancer, but doctors have to review the images many times to identify the 5 percent.

“We hope to use the machine to first pick out the 5 percent, and then the doctors can make the final decision,” said Lin Zhencheng, deputy general manager of Wuhan Landing Medical High-tech Co Ltd.

The machine produced by Lin’s company has an accuracy rate of 85 percent to identify the disease. From March to June, the company helped to finish the examinations of 370,000 women in Central China’s Hubei province.

It was the largest scale of such examinations, and it was almost impossible for doctors to finish the examinations only using manual process.

Chen Kuan said that the hospitals were not only clinical organizations but also academic institutions, and the AI technologies also helped doctors in their research.

Commercialization

According to data from Beijing Infervision, so far there are nearly 200 hospitals using its system and about 50 percent of the chest radiographs were diagnosed using the AI system.

In the beginning, AI was not accepted in the medical sector, but now, many top experts are leading the research of AI, said Ke Yan, a senior executive from Ali Health.

“Only stable AI could be put to clinical use,” said Chen Kuan.

Wang Zhenchang also pointed out that it is very important for the clinical AI products to be simple and easy for doctors to operate.

In the current market, there are many products but they examine the images after the machines take the image, not during the scanning process. He said the future AI technology should be integrated into the machine.

Further, an AI product can only analyze a single disease, and could not match well with the hospital’s work method. A hospital might have to install many AI products.

Data safety is another problem that cannot be ignored, Wang said. Some AI companies set up servers or cloud themselves at hospitals that fall short of national standards on network safety, which carries great risk.

According to data from iyiou.com, by July 31, 2017, China had a total of 131 AI companies engaged in the medical sector.

Both large companies and startups have scaled up the investment and efforts in the field, expecting to gain a share of the medical AI market, but it is not easy.

The biggest problem is that companies in the medical AI industry have not found a mature and consistent business model.

Before the government put into practice the new lists of medical equipment on August 1, the AI medical equipment could not get licenses from the drug authorities, and most products could only be tried for free in medical institutions.

The new measure allows companies to apply for registration of their AI products for medical use, but difficulties still exist.

Companies have to provide detailed information on the products, including test reports, clinic appraisal and risk analysis. After the registration, these products have to go through strict reviews.

Some companies have already found ways to make profit but not from the AI product itself.

Wuhan Landing has cooperated with more than 300 hospitals on its product to detect cervical cancer, making a revenue of several thousands of yuan in 2017, said Lin Zhencheng, deputy general manager of the company.

The company made profit from selling other goods bundled with its AI product used for examinations.

Shang Yang said that in the future AI might be integrated into the existing medical equipment, and AI companies might sell their products or cooperate with the equipment makers who sell the products to hospitals directly.


Related Stories
Share this page
Touched Sympathetic Bored Angry Amused Sad Happy No comment

TV drama director slams practice of fake audience ratingsTrump lashes China over trade during UN speechChina’s sci-tech journals lose appeal for high-quality papers due to management issuesWeChat messages of accuser in sexual harassment case against Chinese billionaire releasedChina denies request for Hong Kong port call by American warshipHow China's private sector will not only survive but thrive amid the trade warNew tariffs take effect as China accuses US of 'economic hegemony' in white paperMattis looks for 'way ahead' after China scraps military talksChina said to cancel trade talks as tariff war escalates: reportNanny responsible for deadly arson case executed
< Prev Next >