Home >

Medical AI "Double Heaven": Giants Compete For Admission To The Bureau, And The Commercialization Of The Card Is Difficult

2020/11/18 11:06:00 0

MedicalAIDual DayGiantCommercializationClinical

The medical AI industry has experienced ups and downs from once hot to gradually in trouble. With the application and verification of artificial intelligence and big data in the epidemic situation, the attention of this industry is heating up again.

Both technology giants and start-ups are rushing into the circuit. For example, Google, Microsoft, Baidu, Alibaba, Tencent have invested a lot of resources in the field of medical AI and launched intensive layout.

Today, although medical AI has been widely recognized as an emerging technology industry, its clinical application and commercialization are still difficult to break through, and the vast majority of medical AI enterprises are still unable to get rid of the fate of "losing money". In addition, there are hidden risks such as reliability and safety in the artificial intelligence medicine which lacks of industry standards. How to put artificial intelligence into clinical use more safely is also a difficult problem to be solved in the industry.

During the epidemic period, the application of medical AI has been further innovation and promotion. Visual China

Problems and risks behind the hot track

It is undeniable that AI and medical technology are still in the initial stage.

No doubt, the "pioneer" of IBM and Watson has been labeled as "pioneer" in the United States, but with the development of "Ai" in the United States, Watson has gradually been labeled as "pioneer".

In 2016, the Medical Research Institute of Tokyo University used IBM's artificial intelligence system "Watson" to diagnose a woman with rare leukemia, which took only 10 minutes. After seeing the hope of AI medical development, IBM put all its money on Watson. In 2017, we invested US $240 million to build mit-ibm Watson artificial intelligence laboratory with MIT. It also plans to invest $3 billion to build Watson's global blueprint.

But before the blueprint was fully implemented, Watson was in trouble. In the last year or two, Watson has been questioned by many industry experts and exposed many problems, including the possibility of prescribing dangerous and wrong cancer treatment plans. In July 2018, the internal IBM documents released by stat, the US health care media, showed that when IBM trained Watson, the treatment plan recommended by IBM for hypothetical patients was based on the program commemorating the experts of Sloan Kettering cancer center, rather than medical guidelines or real evidence.

Zhang Xuegong, Professor of Automation Department of Tsinghua University and part-time professor of life college and medical college, told 21st century economic reporter: "I am not surprised that the development of medical AI similar to Watson is hindered. If there is no particularly excellent technology in the technical level, it is more just hype in the media and propaganda level, so it is difficult to develop."

In addition to Watson, other technology enterprises engaged in medical AI are also facing many industry pain points to be solved. For example, domestic enterprises in China are facing the problem of data flow in medical imaging, and patients are unable to save and manage their original image data.

According to the "2019 China artificial intelligence medical white paper" released by the Artificial Intelligence Research Institute of Shanghai Jiaotong University, China's medical AI is facing challenges in terms of medical talents, data and device approval. Specifically, it includes the lack of medical AI talents, unclear data ownership, inconsistent data standards, high requirements for equipment classification.

During the epidemic period, the application of medical AI has been further innovated and promoted. But it is undeniable that there are various bottlenecks and pain points in the development of medical AI. How to break the bottleneck embarrassment period, and then promote the development of the industry, is placed in front of the medical AI industry is a very key issue.

Zhang Xuegong said: "artificial intelligence contains a lot of challenges. It is not so dramatic that it can solve all medical problems without breaking through one bottleneck. It is a gradual expansion and breakthrough to improve artificial intelligence bit by bit and make it solve some problems in the past in the medical field

Zhu Shunyan, chairman and CEO of Ali health, said at the Guangzhou Cancer Conference on November 14 that the relationship between doctors and artificial intelligence (AI) should be doctor + AI, that is, doctors are in the front, AI is a tool to assist doctors, not the opposite. He pointed out that the most important thing in the medical field is accuracy, which requires doctors to make final judgments and decisions based on a series of indicators calculated by artificial intelligence.

Fan Daiming, an academician of the Chinese Academy of engineering, pointed out that the introduction of artificial intelligence into medicine is an inevitable result. He believes that the best technology is "big data + artificial intelligence". It is inevitable to use it to help medicine to improve human health. "We should learn from it and serve the doctors better." Fan Daiming said.

Although it is considered to be over hyped, it is a recognized fact that medical AI has great potential economic benefits. According to the white paper on artificial intelligence issued by Roland Berger, an international management consulting company, it is estimated that by 2030, artificial intelligence will generate 10 trillion yuan of industrial driving benefits in China. Among them, the medical industry, the use of AI is expected to bring about 400 billion yuan of cost reduction value. As a result, medical AI has become a hot air outlet in the field of artificial intelligence, which is favored by many capitals and attracts numerous players.

At present, the main application scenarios of medical AI are lung CT, fundus screening and medical imaging. In January this year, the joint project team of Nankai University and Beijing inferential technology developed the AI screening system for CT images of new coronal pneumonia in the initial stage of the epidemic. For the first time, the accumulated intelligent recognition technology based on CT images of pneumonia and pulmonary tuberculosis was deployed and applied to Tongji Hospital Affiliated to Tongji Medical College of Huazhong University of science and technology, and Central South Hospital of Wuhan University to assist doctors in rapid diagnosis of new crown pneumonia.

Medical imaging provides image information for medical diagnosis. Zhang Xuegong pointed out that in the specific clinical and commercial applications of medical AI, images are clear and easy to understand, so they are the first to be applied in the field of medical AI. However, there are still limitations. Images can not give all the data in the life system to doctors and patients. Historical data, subjective feelings, and descriptions of various natural languages cannot be separated by images Analysis.

Some rare diseases do not have typical symptoms, and they usually rely on vague description to judge the symptoms. If a large number of cases have been accumulated, it will be more comprehensive to rely on machines to make analysis and judgment than human judgment, because human experience judgment is limited. When it comes to the role of medical workers.

However, at present, AI technology is even applied in some medical fields related to human life safety. AI system also goes deep into diagnosis, prediction and even treatment and rehabilitation. Many new AI tools have been developed, and the corresponding research has been published in some authoritative medical journals. However, due to the uneven quality of experimental design, it is difficult to compare and evaluate the specific effectiveness.

If there is no unified industry evaluation standard, it may bring risks to millions of patients. At the same time, it is easy to encourage medical enterprises to publicize and hype the effectiveness of AI. With the arrival of many giants, AI arena shows great potential, but the rules of the game need to be established.

Deep integration of cross fields or new driving point

With the aging of the population and the growth of patients with chronic diseases, the demand for medical technicians and medical resources is increasing. However, there are still many deficiencies in the existing medical system in dealing with patients who need a large number of long-term diagnosis and treatment and complicated symptoms.

In addition, the high-quality medical resources are unevenly distributed, and there are great differences among regions. According to the report on national medical service and medical quality and safety in 2019 issued by the National Health Commission, the five regions with the highest proportion of patients' outflow are Tibet, Anhui, Inner Mongolia, Hebei and Gansu, while the top five regions with the highest proportion of patients' inflow are Shanghai, Beijing, Jiangsu, Zhejiang and Guangdong. With the rapid growth of medical data, AI big data can save human resources and make up for the shortage of medical labor.

Zhang Xuegong admitted that at present, AI medical really needs to be more applied to the problems that need a lot of repetitive labor to solve. From a longer-term perspective, the development of artificial intelligence should be combined with life research itself.

In recent years, deep learning has become the focus again in the field of AI medicine. Zhang Xuegong said: "the whole field of artificial intelligence is much wider than the field of machine deep learning. Deep learning has its strengths and limitations. We can not put all our hopes on deep learning, but also attach importance to basic knowledge learning. Looking for simple laws from a large number of data and finding complex laws from a small amount of data requires methods other than deep learning or deep learning itself to develop in a better direction. "

According to the report "Research on the application of artificial intelligence in the field of medical and health" recently released by China Development Research Foundation, at this stage, artificial intelligence is widely used in the global medical field, and its popularity is flourishing. The new technology is mainly used in virtual assistant, medical imaging, auxiliary diagnosis and treatment, disease risk prediction, drug mining, health management, medical management, auxiliary medical research platform and other fields.

The report points out that the application of artificial intelligence in the field of medical and health will promote the innovation of health technology and the transformation of medical service mode, promote the reduction of medical cost and the improvement of medical service efficiency; at the same time, it will also help to form a homogeneous, standard and easy to extend medical service system, optimize the allocation of resources, and ensure the demand side, especially for people in remote areas, to enjoy by everyone The right to high quality and high standards of health care, and to promote health equity and accessibility.

In the future, AI medicine will be extended to almost every field and category. From medical devices, surgical equipment, all kinds of passive implants, such as artificial joints, artificial organs, cardiovascular stents, etc., AI medical robots will emerge as the times require. The interdisciplinary team construction is needed for the deep integration of AI and medical devices, including personnel training and deep integration research. The development of "Ai" and "quasi medical" will be fast. Enterprises should not meet the current needs, and should be prepared for the future development of smart medical high-speed iteration.

 

  • Related reading

"Digital Medical Economy" Of Multinational Pharmaceutical Enterprises: Omni Channel Layout And Intelligent Course Management

Professional market
|
2020/11/11 20:04:00
1

Acceleration Of Digital Medical Layout Of Multinational Pharmaceutical Enterprises In China After Epidemic

Professional market
|
2020/11/11 20:04:00
0

Bid Farewell To Shenzhen Real Estate, Evergrande Seeks To Split And Go Public To Reduce Debt

Professional market
|
2020/11/10 12:34:00
0

Yuantong'S "Blocking" The Polar Rabbit: Its 144 Branches Violate The Regulations To Act As The Agent Of The Pole Rabbit Business

Professional market
|
2020/11/6 13:30:00
196

The Goal Of New Energy Vehicle Industry Planning In 2025: From 25% Back To 20%, Power Consumption Reduction And Charging Network Layout To A New Height

Professional market
|
2020/11/3 20:01:00
194
Read the next article

China Textile Federation Environmental Protection And Resources Commission Is Honored As "WWF Green Transformation Pioneer"

Recently, a series of activities commemorating the 40th anniversary of WWF in China were held at