The recent matter about vaccines, we can call it a crisis, and even some people will link the previous tainted milk powder incident together.
In this matter, the conscience of the fraudulent enterprises is greatly bad, but the medical supervision system has a major problem, which is also a problem that we cannot avoid.
What should we think about behind the vaccine problem? What progress can artificial intelligence bring to the medical industry?
Sanofi Pasteur uses AI to develop influenza vaccine
Sanofi Pasteur is a world-renowned R&D company. Recently, it has reached a cooperation with a biomedical company called BERG to use artificial intelligence to develop influenza vaccines.
BERG's platform uses its developed artificial skills technology to perform data-driven analysis of high-throughput molecular and clinical information. It can extract actionable insights from scattered, huge data, and this method has also been proven.
Using this artificial intelligence platform, Sanofi Pasteur will look for potential biomarkers that can evaluate the effectiveness of seasonal influenza vaccines. If all goes well, these markers can predict the breadth and durability of the immune response caused by the flu vaccine, which will help future vaccine development and benefit patients.
Disease diagnosisArtificial intelligence can diagnose colorectal cancer in less than 1 second, with an accuracy rate of 86%
Diagnosing cancer correctly and in time has always been a problem. Recently, a group of scientists from Japan used artificial intelligence to make a diagnosis of colorectal cancer in less than one second, with an accuracy rate of 86%. What's amazing is that it can make a diagnosis even before a benign tumor worsens.
In this study, scientists let artificial intelligence conduct in-depth observations of polyps in the colorectal. They magnified the polyp 500 times so that the artificial intelligence can carefully see the changes in these tissues. Subsequently, they provided 30,000 pictures of cells before and after cancer, and used machine learning methods to train artificial intelligence. In the end, this system can make a diagnosis in just one second, with an accuracy rate of 86%. This is also the first time that artificial intelligence has been used in the diagnosis of colorectal cancer.
Diagnose breast cancer, the accuracy rate of artificial intelligence is as high as 97%
In the United States, 40,000 women die of breast cancer each year. This regrettable result is related to the lack of timely diagnosis of breast cancer. According to statistics, if breast cancer can be detected at an early stage, they may even be cured.
On the other hand, breast X-ray examination, which is a routine examination method for breast cancer, has the problem of a high false positive rate. Some patients will have suspicious tissues in their breasts under X-rays, and they will also choose surgery to remove them. However, analysis of these tissues after surgery often finds that they are benign. In other words, these women received unnecessary treatment. So, is there any technology that can reduce the false positive rate while ensuring the diagnosis of breast cancer?
Artificial intelligence can do it. The system developed by researchers at Harvard Medical School and Massachusetts General Hospital can make a diagnosis from a series of data points. In addition to biopsy results and pathology reports, this artificial intelligence will also analyze the patient's family medical history and ethnic information. In this way, the accuracy of diagnosis is greatly improved.
Among 335 high-risk pathologies, the artificial intelligence has a diagnosis accuracy rate of 97%. Researchers say that because of its accuracy, the probability of patients undergoing unnecessary surgery has dropped by 30%! For patients, this is a great boon.
There is no doubt that the era of artificial intelligence has already arrived. Researchers who know how to use artificial intelligence have a significant advantage over those who do not know how to use artificial intelligence.
Based on previous research results, Yiou Think Tank released the 2018 China Medical Artificial Intelligence Development Research Report.
From the perspective of the development of medical artificial intelligence, the report summarizes 10 major insights into the development of medical artificial intelligence in China, taking commercial implementation as the starting point. It sorted out 10 mainstream medical AI products in China and the top ten mainstream products in the field of medical artificial intelligence, and analyzed it from the perspectives of technology maturity, use effect, development, and corporate cases. The following is the full text of the report:
2018 China's top ten insights on medical artificial intelligenceFrom the perspective of the four core application scenarios of artificial intelligence in the medical and health field-medical imaging, virtual assistants, health management, and drug development, ten insights and related views on the development of medical artificial intelligence in China are put forward.
1. Some intelligent imaging diagnosis companies will obtain Class III device certificates in 2018 and officially enter the commercialization stage.
2. The competitive landscape of intelligent imaging diagnosis has basically taken shape, "pseudo-medical AI companies" have basically been eliminated, new technical players have basically no possibility of obtaining venture capital, and business opportunities have already been missed.
3. Voice electronic medical records: The cost of landing in the hospital is high, the products need to be customized in the department, and the customer unit price is low. It is mainly used in pathology and imaging departments.
4. Intelligent consultation: The construction of knowledge map is the key. Currently, it only plays the role of guiding diagnosis, assisting retrieval or connecting doctors and patients. The in-hospital scene "pre-interrogation" is in high demand and has the ability to land.
5. Chinese people's awareness of health management has yet to be cultivated, and big health data has yet to be collected and integrated. Enterprises take the B-end as the main entry point.
6. There is a serious shortage of psychotherapists, and AI may become an alternative tool.
7. The quality of compound data in drug development is the key to AI companies.
8. With the help of international forces, Chinese AI drug R&D companies have emerged from nothing, and it is expected that more players will emerge from 2018. AI drug R&D may be a new frontier in the future.
9. The product form is mainly software/SaaS, and the business model of charging software license fees has certain limitations. The commercial landing of software and hardware integrated products has more advantages.
10. The overall amount of medical data in China is large, but the amount and quality of data for segmented scenarios still cannot meet the training needs of algorithm models; the lack of follow-up data makes domestic research in areas such as "five-year survival rate of cancer patients" blank.
In order to interpret the status quo of commercial implementation in a more in-depth manner, in the report, Yiou think tank analyzed and evaluated the top ten medical artificial intelligence products mainly according to the two dimensions of technology maturity and use effect. Among them, the two dimensions of technology maturity and use effect are mainly judged through the product appearance time, landing situation, development situation, number of enterprises, industry professionals and expert interviews. In addition, it also analyzes in more detail from the perspectives of product development and enterprise cases involved.
Six development trends of medical artificial intelligenceCombining policies and the current status of commercial products, Yiou think tank believes that the market will present six major trends this year:
1. Starting in 2018, the landing speed of AI imaging products will accelerate, and the maturity of product performance will continue to improve.
2. With the improvement of technology maturity and the acceleration of the penetration rate of voice electronic medical records in hospitals, leading enterprises can form economies of scale.
3. With the continuous improvement of the knowledge map, the pre-inquiry function can effectively improve the efficiency of doctors.
4. The development of health big data will further increase the application of AI in health management scenarios.
5. AI will penetrate deeper into mental and psychological health, and it may become the core driving force in this field in the future.
6. The field of AI + drug research and development will give birth to unicorns.
Four major challenges in the development of medical artificial intelligenceOne is the problem of data quantity: the overall amount of data in China's medical care is large, but the amount and quality of data for different diseases are uneven, and training data for some diseases is lacking; the problem of healthy big data islands has been alleviated, but deep learning has not yet reached The stage.
The second is the problem of data quality: the accuracy of labeling in AI data processing is related to the accuracy of the results, and a large number of doctors are still required to label in the past two years. The quality of data in drug research and development is critical to the improvement of research and development efficiency.
The third is the issue of talents: AI algorithm talents and medical talents have different knowledge systems. How to integrate their respective advantages to maximize value is worth thinking about.
The fourth is the market problem: medical care is considered to be the first field of artificial intelligence, but the particularity of medical care will have higher requirements for products, from the recognition of acceptance to the improvement of the corresponding payment system and the access to medical insurance. It takes a long process.
SHENZHEN CHONDEKUAI TECHNOLOGY CO.LTD , https://www.szsiheyi.com