To sort out the hot words of the medical device industry in 2018, "Artificial Intelligence (AI)" is one of them.Regardless of China International Medical Devices (Spring) Expo, or the first China International Import Expo, AI is a hot spot in the highlights of the show and is sought after by all parties.However, it is undeniable that the application and promotion of AI medical products face many practical problems.
Recently, the China Biomedical Engineering Society held the "Medical Artificial Intelligence Data Governance and Standardization Conference" held in Beijing.In addition to the prospects for the application of AI, the guests also made more rational analysis and discussion on practical issues such as laws and regulations, standards system, and quality evaluation.
“AI will cause major changes in traditional medical diagnosis, medical services and health management. To adapt to these changes, close coordination of policies, regulations and regulatory mechanisms is required.”Prof. Zhang Wei, academician of the Chinese Academy of Sciences and dean of the Institute of Artificial Intelligence of Tsinghua University, pointed out in his speech that it is necessary to re-establish a system of human-machine collaborative medical care with mutual trust, reliability, high quality and high efficiency.
Present situation
There is still a gap between performance and human level
When the AI was born in 1956, a symbolic reasoning model based on knowledge and experience was proposed, and medical diagnosis was the earliest application field of this model.
In the early 1970s, in order to regulate the use of antibiotics, the research team at Stanford University in the United States spent five or six years developing the MYCIN system.In the 1980s, some clinical institutions in Beijing also developed computer diagnostic systems for diseases such as hepatitis and acute abdomen.However, these systems have not been widely used. What is going on?
"The MYCIN system failed to apply because there is no legal basis for the machine to see a doctor.In addition, from a technical point of view, some diagnostic systems introduced in the 1980s are based on knowledge-driven methods. Medical knowledge comes from experts. Because expert knowledge is difficult to obtain, and it is scarce and expensive, it is difficult to construct practical application scales system at that time. Zhang Wei explained.
With the popularity and application of the Internet, there has been unprecedented big data in the medical field. The deep learning model based on big data in AI came into being and was widely used.However, in some performances, AI medical applications are still far from human levels.
Zhang Wei pointed out that the deep learning-based system has unexplained and fragile reasons, and it is difficult to obtain the trust of clinicians, thus hindering its widespread application.At the same time, based on deep learning of big data, the quality of the data usually used is not high, besides, the "black box" learning method.Although the AI system outperforms humans in terms of individual indicators, it is far from the level of human beings in terms of other properties such as robustness and interpretability.
"In order to solve these problems, we need to strengthen research on the one hand; on the other hand, we need to develop a complete set of laws and regulations, standards systems and regulatory measures." Zhang Wei said.
"AI medical products will present new challenges to the existing medical legal system.Regardless of risk prevention, code of conduct, or legal subjects, while promoting the development of AI medical products, it is necessary to conduct in-depth discussions on its possible risks and legal liability.”Cao Yanlin, director of the Institute of Medical Information at the Chinese Academy of Medical Sciences, said that in 2016, AlphaGo's victory over Li Shishi meant that human beings fell in a chess game that was considered an intellectual bastion.Under the catalysis of this event, it has once again aroused widespread and sustained attention to AI in all walks of life.Especially in the past two years, with the development of technology industries such as Baidu, Tencent, and Keda Xunfei in the AI field, AI has become the focus of government and scientific research institutions.
Make progress
Standardization and construction work is well organized
"AI medical equipment enters the critical period of product transformation in China, and AI product quality evaluation standards and specifications are rapidly being researched and established."Ren Haiping, director of the China National Institute for Food and Drug Control (hereinafter referred to as "the Central Inspection Institute"), told reporters.
Zhong Daidi, a professor at Chongqing University and chairman of the International Personal Health Equipment Standards Committee, said that due to the special nature of medical devices,AI products and technology applications in the medical field must operate within the industry policy framework for medical device supervision and management.Follow the requirements of the medical device industry for safety, effectiveness, ease of use, etc., and these need to be implemented through technical standards and supporting policies.
Zhong Daidi further said that on a global scale, the traditional "decision-support" product division can no longer meet the needs of industrial development.The relevant standards for AI products in the medical fields of all countries are in a blank state, and it is urgent to carry out a special standard system construction to provide support for all stakeholders.Standardization work has a basic, supportive and leading role in the development of medical AI industry.
The reporter learned that as a national medical device testing center, the Central Inspection Institute began to lay out the quality evaluation research of AI medical devices in 2016.At present, the Chinese Procuratorate has undertaken the registration inspection of dozens of AI medical device products nationwide. On this basis, in September 2018, the State Drug Administration approved the Central Inspection Institute as the national standardization management unit for artificial intelligence medical devices, and the preparation of the expert group is about to be completed.In December 2018, the Chinese Inspectorate was approved by the Institute of Electrical and Electronics Engineers (IEEE) Standards Committee as the convening unit of the Artificial Intelligence Medical Device Working Group, and two international standard project proposals have been approved.
"China's AI medical device standardization work will be launched at the same time on the domestic and international stage, and will certainly support and guide the review and approval and industry development." Ren Haiping said.
Prospects
Medical devices ushered in a period of innovation opportunities
Despite real problems and challenges, AI has a lot to offer in the medical field, which is widely recognized by experts and business people.
Cao Yanlin pointed out that AI Medical can not only help patients to conduct preliminary analysis and assessment of self-health status before the visit, assist doctors in managing patient information and improve service levels,but also can help hospitals guide and manage patients' medical treatment, conduct drug and medical device research and development, alleviate the shortage of medical human resources, reconstruct medical service models, and eventually reconstruct the health care system.
According to Zhao Yiwu, chairman of Beijing Natong Technology Group Co., Ltd., compared with the field of medical diagnosis and analysis, AI has been developing in a period of treatment, especially in surgery, due to factors such as safety and ethics.However, in the past year, with the in-depth development of algorithm research and the expansion of applicable fields, the application of AI in surgery has spawned the concept of precision surgery, which is also a powerful supplement to precision medicine.
He said that in the face of new market space and new areas, whether it can change research and development ideas, strengthen industry understanding, standardize medical behavior, and encourage cross-domain cooperation,opening up a new world in the field of surgery is the biggest challenge and opportunity for AI healthcare companies.
Associate Professor Wang Defeng of the Chinese University of Hong Kong also said that with the release of the "New Generation Artificial Intelligence Development Plan" and "Made in China 2025", the innovation and research and development of high-performance medical equipment has been pushed to an unprecedented height.
"The future of the entire medical device must be intelligent, and R & D investment will also develop in this direction. Medical robots are the ultimate direction of medical device intelligence, and the proportion in the medical device market will also grow.”Wang Defeng further said that AI combined with robots must analyze specific problems to form a "specialist customized" medical robot.The application of AI technology to multi-modal and high-precision automatic fusion of image data is an important cornerstone of the combination of frontier technology and intelligent robot application scenarios. (Reporter Ma Yanhong)
p>ht:normal;text-transform:none;font-style:normal;font-size:10.5000pt;mso-font-kerning:0.0000pt;" >