Thanks for reading the 118th article of Tencent AI Lab WeChat ID. This article will introduce "AI+Medical/Medicine" to help blood disease screening and in vitro diagnosis.
Shenzhen, January 29, 2021-Tencent AI Lab announced that it has signed an AI cooperation framework agreement with Shenzhen Mindray Biomedical Electronics Co., Ltd. (hereinafter referred to as Mindray Medical) to jointly develop AI products for blood cell analysis and further explore AI technology in vitro diagnostics Integration and application of The cooperation will rely on Tencent Miying·Open Laboratory to accelerate the construction of a new generation of artificial intelligence open innovation platform for medical imaging of the Ministry of Science and Technology, and provide the public with better quality medical equipment and services.
Through this cooperation, the two parties will further develop blood cell morphology testing equipment and system applications with world-class imaging quality and cell recognition capabilities. After large-scale production in the future, it is expected to relieve the work pressure of primary endoscopy physicians and improve the diagnosis and diagnosis of blood diseases. The identification level promotes the development of blood cell analysis technology in China.
Yang Wei, General Manager of Tencent AI Lab, said: "AI+Medical/Pharmaceutical" is one of Tencent's AI Lab's long-term core research directions. We are very happy to cooperate with industry top partners like Mindray to enable technology to help medical care, assist doctors, and serve patients .
Dr. Tianyi Qian, general manager of Tencent Miying, said that relying on Tencent Miying·Open Lab, Tencent will further open up medical AI capabilities and technologies to the entire industry, covering five modules of application center, cooperation center, algorithm center, data center, and management center to help The industry solves the problem of popularization of medical AI and incubates innovative applications of medical AI in batches.
Regarding this contract cooperation, Yan Huawen, General Manager of Mindray’s IVD Division, delivered a speech. He said: “With Mindray’s deep understanding of clinical needs and accumulation of detection technology over the years, and Tencent’s leading exploration in artificial intelligence technology, The results of the cooperation between the two parties will greatly improve the ability of clinical intelligent detection."
Li Chaoyang, general manager of Mindray Technology Research Institute, said: "Such cooperation will surely promote the return of the most important "reading" link in blood cell analysis to the clinic, and improve China and the world. The high level of blood cell analysis quality has enabled more primary hospitals to have accurate and efficient screening capabilities, which can truly help the early diagnosis and treatment of leukemia and other diseases, so that high-end medical technology will ultimately benefit everyone."
Solve the clinical pain points of insufficient manpower for microscopy
Morphological analysis of blood cells, also known as routine blood tests and blood tests, is one of the most widely developed clinical testing items. From infectious diseases such as pneumonia to bleeding diseases such as leukemia, blood cell analysis can provide important information for the diagnosis and identification of blood diseases. information.
Take leukemia as an example. The incidence of leukemia in China is about 2.76 per 100,000. It is the sixth most common tumor in China, and it is also the most common tumor in children and adolescents. It has the characteristics of high malignancy and rapid development. With the development of medicine, the cure of leukemia is gradually being achieved, and the complete remission rate of adult acute leukemia has reached 60%-80%. If childhood leukemia can be detected in time and appropriate treatment methods, about half of them can survive or be cured for a long time. Blood cell analysis is the first screening tool for patients to see a doctor. Early leukemia screening can be achieved through morphological testing, prompting early clinical intervention and improving the cure rate.
For a long time, blood cell analysis has relied on examining physicians to observe blood smears with a microscope. Morphological examination is a highly empirical task, but it is tedious and time-consuming. In recent years, with the gradual improvement of medical conditions, the number of samples in hospitals has increased exponentially. However, not only has the number of high-quality microscopy talents not increased accordingly, but because of high requirements, slow training, and occupational diseases, the group of young doctors has become engaged in microscopy. The willingness of inspection is low, and there is a crisis of "diagnosis" for microscopy talents.
On the one hand, it is difficult to train morphological examiners and the cycle is long. On the other hand, the number of laboratory samples is increasing, resulting in a serious shortage of examiners. Therefore, the proportion of morphological examinations carried out in many hospitals is obviously low. According to the recommendations of ICSH (International Committee for Hematology Standards), the ratio of morphological examinations in general hospitals should reach 20%-30%, and research shows that the actual microscopic examination rate of most laboratory departments does not reach the above recommended ratio, and some hospitals may be less than 5 %, not even microscopy. This may lead to risks such as missed inspections, and the resulting medical disputes are also reported from time to time.
In response to this situation, many domestic and foreign companies have launched a variety of morphological analysis automation equipment, but they all have the problem of insufficient recognition accuracy, and still need a lot of manpower to review and confirm, so they have not fundamentally solved the clinical pain points.
AI helps healthcare move towards intelligence
In this cooperation, Tencent AI Lab will use machine learning, computer vision and other cutting-edge algorithms and the rich experience of top research scientists, Tencent Cloud’s comprehensive solutions, combined with Mindray Medical’s expertise in slide processing, imaging systems, integrated design, and clinical experts , Massive data and other aspects, jointly develop an automated blood cell analyzer to further improve the accuracy of recognition, and is expected to help solve the problem of "lack of people" in domestic hospitals and "lack of experience" in doctors.
This automated blood cell analyzer can detect cell quality changes, and its cell identification capability can reach the world's leading level. At the same time, it has a cost advantage compared to foreign products with high unit prices.
In terms of blood cell morphology analysis, Tencent AI Lab uses a number of self-developed Auto-ML, Auto-Augmentation, and NAS advanced AI computing solutions to face huge amounts of data mixed with noise, greatly reducing training time and increasing the number of iterations . At the same time, algorithm engineers use multi-agent simulation and a decision-making system composed of multiple expert systems to improve the analysis performance of the overall cell morphology, so that abnormal and malignant cells are not missed, and suspected cells are prompted.
In terms of equipment manufacturing, Tencent AI Lab uses technical solutions such as model compression and computational graph simplification to help products achieve greater advantages in computing efficiency and cost optimization.
In addition to morphological testing, Tencent AI Lab will also work with Mindray Medical to explore the integration and application of artificial intelligence technology in other areas of in vitro diagnostics, such as urine analysis, smart laboratory, and in vitro diagnostic big data, to provide the public with better quality Medical equipment and services to help the development of medical technology in China and the world.
Tencent AI Lab technology covers the entire diagnosis and treatment process
"AI+Medical/Medicine" is one of the core research directions of Tencent AI Lab, which is also a research topic that is expected to benefit all mankind. Especially in the context of the new crown epidemic that has ravaged the world for a whole year, AI technology has unparalleled application prospects.
Generally speaking, treatment is divided into three stages: pre-diagnosis, in-diagnosis, and post-diagnosis, involving various information exchanges and complex decision-making between patients and medical staff. Part of the work is highly repetitive, time-consuming, and labor-intensive, which will squeeze the doctor’s Really effective diagnosis time. AI can help doctors improve efficiency and accuracy, and is expected to alleviate the imbalance between doctors' supply and demand.
Tencent AI Lab continues to promote AI to cover the entire diagnosis and treatment process. The medical team comes from the National Institutes of Health, MD Anderson Cancer Center, Stanford University, Siemens Medical and other top international institutions. Since 2017, the laboratory has given full play to the advantages of cutting-edge algorithm technology and top talents, and has possessed a series of cutting-edge technological capabilities such as auxiliary diagnosis, image analysis, pathological analysis, and new drug research and development. It has continued to provide a national artificial intelligence open innovation platform for medical imaging built by Tencent. And Tencent Miying and other products provide technical support, while continuous breakthroughs in independent technology research and development and application.
In terms of research, Tencent AI Lab's related research has repeatedly appeared in top academic journals and conferences, such as Nature Communications , the top medical imaging conference MICCAI , and the world's top radiology conference RSNA AI Challenge . During the epidemic, with the team of Academician Zhong Nanshan, the AI prediction model for the probability of COVID-19 patients developing to a critical condition was jointly released, which was also included in the Nature sub-Journal .
In terms of application, Tencent AI Lab, Sunny Optical Technology, and Jinyu Medical jointly developed the intelligent microscope for pathological analysis and obtained the NMPA registration certificate in April 2020 , becoming the first relevant product approved for clinical use in China; July 2020 May launched the "Cloudland Chi medicine" platform, covering five modules of preclinical drug discovery process; in 2020 November with the domestic first listed company in the field of pathology An Biping announced to build a liquid-based cervical cytology AI database.
* Welcome to reprint, please indicate from Tencent AI Lab WeChat (tencent_ailab)