In recent years, AI has gradually penetrated the medical profession. A number of top three hospitals in Guangdong are in the process of laying out AI medical care. At the same time, Internet companies such as BAT, "big coffee" are also rushing into the AI medical field. On July 4 this year, the Guangdong Provincial Health Planning Commission held the kick-off meeting of the “Guangdong Province Internet + Medical Health Development Action Plan”, clarifying that the Provincial Health Planning Commission will implement eight major actions such as “building a telemedicine service system” in the near future.
AI helps doctors "watch movies" in a matter of seconds, improving efficiency and reducing misdiagnosis. It can also be used as a "doctor's assistant" to deepen the knowledge and experience of medical experts and reduce the differences in doctors' knowledge... AI is useful in the medical community. Wuzhidi, although the domestic market is at an early stage, there are still a number of high-quality medical big data companies such as Carbon Cloud Intelligence, Zero Technology, Medical Duyun, and Pushu Technology, which are developing rapidly. However, there are still many difficulties in developing AI medical care. The premise of the above scenarios is that it must be based on the full collection, storage and processing capabilities of medical big data, as well as high-quality data polished by structured processing. In a survey interview, Nanfang Daily reporter found that this is also a problem faced by many AI medical companies, and how the core competitiveness is seen here.
Business case
Medical big data platform to help large sample research
At the 2018 American Society of Clinical Oncology (ASCO) meeting held in June this year, a large sample of the team from Professor Liang Wenhua from the First Affiliated Hospital of Guangzhou Medical University/National Clinical Research Center for Respiratory Diseases won the annual meeting. The big prize "Merit Award".
It is understood that this study analyzed the relationship between non-small cell lung cancer (NSCLC)-driven genes and chemotherapy/PD-L1 (programmed death ligand) blocking therapeutic sensitivity markers, and the results explained patients from the perspective of genetic mutations. The reason for the different sensitivity of different drugs, thus guiding patients to use more suitable and effective drugs for treatment. Patient cases are from China, and PD-L1 inhibitors are currently popular anti-tumor drugs.
In the above study, LinkDoc's medical big data platform played an important role. As the first enterprise in China's medical big data and artificial intelligence field with a valuation of more than 1 billion US dollars, Zero Technology adopts the model of “data + technology + platform + service” for medical institutions, industry regulatory authorities, governments at all levels, and the pharmaceutical industry. Insurance companies and other providers provide big data overall solutions. At present, the company's product line includes Hubble (AI assisted decision system), research project management system and so on. As of the first half of 2018, the penetration rate of single tumors in the Zero-Year Medical Big Data platform has reached 60%, surpassing the United States.
Through the LinkDoc Hospital Electronic Medical Record System (EMR), the scientific research personnel of the First Affiliated Hospital of Guangzhou Medical University realized the structured data entry, and a professional quality control team conducted data verification. Finally, the professional statistical team conducted statistical analysis of the data. Because the data processing process is completely completed by the data model, human intervention and various types of bias are greatly reduced, and the research results are as objective and accurate as possible.
It is worth noting that statistical analysis and research of data is a huge project. If you manually obtain data by browsing the form in the past, it is difficult to complete without more than one year, and there are many research hypotheses and research methods. Difficulties. But with the support of the medical big data platform, it took only three months to complete the relevant work, which is a major advancement in big data in the medical field.
Technical Difficulties
Tumor database structuring is the key
How does the medical big data platform help large sample research? In this regard, according to the relevant person in charge of Zero Technology, Zero Technology has spent three years working with 700 comprehensive and specialist top three hospitals nationwide to help more through the analysis and replication of these cases. The patient treats the tumor.
“We have established a LinkDoc real-world tumor database covering more than 50 tumors, covering all clinical modules. At the same time, we have developed the world's first medical record structured artificial intelligence DRESS engine with independent intellectual property rights.” For the establishment of a tumor database, The person in charge said: "Structuralization is the most critical and very difficult thing."
He Guowei, CEO of Philips Greater China, also told the media that in 2020, China's data volume will account for 20% of the world's total, becoming the world's largest data resource country. There is no shortage of medical health data in China, but the data is “mistaken and misjudged” by various reasons. It is unstructured and cannot be used directly, which poses challenges for the development of AI.
Professor Wu Xiaojian, deputy dean of the Sixth Affiliated Hospital of Sun Yat-sen University and director of the five districts of the Department of Colorectal and Anal Surgery, told the Nanfang Daily reporter that the application of artificial intelligence technology in the medical field will help promote the establishment of our own Chinese disease database and establish high quality The Chinese population genetic database will provide an important pillar for China's complex diseases, cancer treatment and drug research and development. It is understood that Zhongshan Sixth Hospital has cooperated with IBM Watson to establish the Watson Gastrointestinal Disease Artificial Intelligence Medical Center. It is a national leader in the sixth hospital affiliated to Sun Yat-sen University in the diagnosis and treatment of gastrointestinal cancer and inflammation, and has considerable patient data, especially Single disease data.
Development proposals
Guide enterprises to break through technical barriers
Artificial intelligence can help and assist doctors to grasp patients more accurately. Although there are still many needs to be improved and improved at this stage, it will definitely be the future development trend. Many AI medical companies have expressed the expectation that the government and associations can play a leading role in promoting industrial development.
"If the team only has technical genes, it is often unclear what the industry needs are. It is very difficult to make industrial transformation. The medical industry is a strong business-oriented industry. The business accounts for 60% to 70%, and technology accounts for 30% to 40%. The Internet often says 'trial and error', but The underlying logic of medicine can't be proved without being able to do it. It can't take patients to 'try the wrong'. It is very necessary for the government, associations, etc. to pull."
It is understood that the demand for artificial intelligence in the medical field is very urgent. However, the current medical AI field is not fully mature technology, and it can be said that it is still in the "infancy period." For different patients and their appearance, even if the symptoms are the same on the surface, the final treatment plan may be completely different, which is currently not fully realized by AI products.
According to Zero Technology, the current development bottleneck is that AI is still learning through a model. It has not yet reached a completely personalized distinction in the medical field, and it cannot replace the communication between people. Therefore, it is hoped that in the future, the government and industry associations will introduce more policies that are conducive to the development of artificial intelligence, introduce more such enterprises, promote exchanges and cooperation between enterprises, and jointly promote the development of artificial intelligence industry.
Wu Xiaojian said that the country has just begun to pay attention to data, and there is still a development process. At present, AI is not mature in treatment decision-making, but no matter how mature AI technology is, patients still need emotional attention and interaction. AI can help doctors do what they are better at, rather than replace.
Stone of other mountains
Mayo Clinic:
Use AI to predict the probability of illness
How to use artificial intelligence and big data to improve medical services, the global medical institutions are facing the same problem. How did the Mayo Clinic, known as the “Medical Mecca”, respond?
At the second Mayo Clinic China Hospital Management Summit held recently, Mayo Medical Group's Chief Technology Officer Steven J. Demuth gave a keynote speech on "The Application of Artificial Intelligence and Medical Big Data in Medical Services."
He said that Mayo will record the events in the hospital and generate digital portraits of hospitals and patients. This allows you to simulate what will happen in the future to better manage the interaction between the hospital and the patient. Using big data and optimizing data will help us predict what will happen in the future. After collecting data, we can analyze who is more likely to get sick, how they are sick, and how we should treat them.
It is worth noting that not only Mayo, but also the application front-end of AI applications in many top hospitals in the United States, including Johns Hopkins Hospital, Massachusetts General Hospital, and University of California Los Angeles (UCLA) Medical Center.
In general, the medical AI application segment is divided into disease prediction, hospital management, assisted diagnosis, precision surgery and health management. Unlike domestic medical AI, which focuses on assisted care, US major hospitals focus on pre-positions such as disease prediction and health management in the context of medical AI applications. (Reporter Li Wei)
0 comments:
Post a Comment