Lian Ying Shen Dinggang: AI is used in many medical scenarios, and talent training in cross-disciplinary fields is the key

Cutting-edge technology can benefit the grassroots

"A doctor who understands AI can replace a doctor who does not understand AI. A doctor who knows AI can replace an AI who does not understand a doctor."

At the main forum of the 2018 Medical Technology World Forum, Professor Shen Dinggang, the joint CEO of Lian Ying Intelligence, once again voiced medical artificial intelligence.

For a long time, domestic medical imaging hardware equipment and infrastructure resources have been unevenly distributed. In order to solve this dilemma, the state implements a grading diagnosis and treatment policy. However, this policy faces various problems in its implementation. The lack of talents, especially high-quality doctors, is one of its core pain points.

Professor Shen pointed out at the meeting that the use of AI to assist doctors in diagnosis, especially to achieve expert-level AI, can effectively improve the diagnostic level of primary hospitals, thereby alleviating the scarcity of medical personnel and promoting the implementation of graded diagnosis and treatment.

He said that the parent company affiliated with Lian Ying Intelligence, as a high-end medical equipment company, has launched 56 products in the past, installed more than 4,300 sets across the country, and is rooted in hospitals across the country in a similar way to nodes. Professor Shen Dinggang said: “In a region, all provinces, counties, and township hospitals can realize the cloud connection of different hospital imaging equipment through image cloud, so that hospitals at all levels form an image center with interconnection and resource sharing. The AI ​​smart application we developed to assist doctors in intelligent diagnosis and early screening can be shared with grassroots hospitals through the cloud to help grassroots doctors achieve smarter and more accurate diagnosis, improve the level of primary hospitals, and relieve radiologists to a certain extent. The problem."

Medical artificial intelligence requires high-quality talents in the cross-disciplinary field who understand both doctors and AI

Medical artificial intelligence is very different from "AI+ industry" and "AI+ smart driving". This difference is not only reflected in the differences in application scenarios, but also in the difference between data and algorithms. To control "AI+Medical", you need an excellent cross-disciplinary talent who understands both doctors and AI.

Professor Shen said at the meeting: "Enterprises must take root in the field of medical artificial intelligence, and it is essential to have experienced talents. In particular, talents who have worked in the industrial industry for many years and have rooted in the academic field for many years. Now, the company has already owned A group of industrial elites from Apple, Google and Tesla, as well as associate professors from Cornell University, and so on, have such a series of talents that we can make full-stack medical artificial intelligence."

"But we also need a large number of young talents to join us." Professor Shen said. Lian Ying Intelligent pays special attention to the cultivation of AI talents. As early as June of this year, Lian Ying Intelligent took the lead in making action in the direction of AI talent training - the establishment of the Joint Imaging Intelligent Medical Intelligence Research and Cultivation Research Center. “At this center, we hope to let science and engineering researchers enter doctors, service doctors, and empowerment doctors. At the same time, doctors, especially young doctors and science and engineering researchers, can grow together. Because for doctors, AI does not It may replace a doctor, but a doctor who knows AI can replace a doctor who does not understand AI. Under the guidance of senior scientists, young doctors and science and engineering personnel will naturally become a very good partner to bring medical and engineering at a faster rate. Combine."

AI can be applied to the vast medical scene

The application scenarios of AI are very extensive. Professor Shen gave a detailed introduction to the AI ​​application developed by Lian Ying Intelligent in his speech.

“In the case of lung cancer, artificial intelligence can assist imaging, lung cancer screening, and follow-up. Doctors can compare the latest patient lung nodule images with historical images to diagnose patients with problems and At the same time, doctors can use artificial intelligence to make prognosis and prediction during the treatment process. In these processes, artificial intelligence can greatly improve the efficiency of doctors. In general, from screening, follow-up, diagnosis, treatment to prognosis, These can be optimized by artificial intelligence, and the joint intelligence is the full-stack artificial intelligence for the entire process."

“Artificial intelligence can also empower devices. In MR and CT, doctors can perform a one-button smart scan to get a three-dimensional image of the patient. The patient lies on the scanner and the AI ​​identifies where the organ to be scanned is. This is the computer. Visual application."

"Artificial intelligence can also do a series of auxiliary diagnosis. For example, AI can automatically judge more than ten kinds of lung diseases according to X-ray chest; it can detect the location and parameters of tumors; it can be used for bone injury identification based on CT images. These applications are very urgent in emergency department. In addition, doctors can understand the flow of blood in the CT image, especially the flow of blood in the heart, to determine what kind of stent the patient needs, and the subsequent improvement in the success rate of the operation. At the time, the doctor usually takes 20-30 minutes to outline the organs inside. Now, the joint intelligence can automatically outline an organ in 0.7 seconds, and quickly complete the organ segmentation."

“Artificial intelligence also enables automatic parameter detection of arthritis. All of these AI applications can be placed in the image cloud for assisted analysis by telemedicine .”

In order to describe the application of artificial intelligence in depth, Professor Shen showed the following cases to the audience in his speech.

“Artificial intelligence can make intelligent assessment of brain structure. Doctors can obtain the diagnosis results of mild cognitive impairment and Alzheimer's disease through AI. For example, by analyzing the brain images of patients 60 years old, 61 years old, 62 years old by AI, we can get Knowing the corresponding changes in each area of ​​the patient's brain, and then generating a structured report, assisting the doctor to make a very accurate diagnosis for the patient."

"In addition, AI can also do intelligent scanning of tumors. When a doctor thinks that a patient has a tumor, he can make a precise scan of the patient. After the first image is scanned, the AI ​​method can determine the approximate tumor type. Then decide what to scan next. Then, the AI ​​combines the image of the subsequent scan with the previous image to continue to judge what the third image scans. The image scanned by this form has a better diagnosis."

"In terms of chest radiographs, in addition to regular readings, AI can also be used for the re-reading of X-ray chest radiographs. Every night the doctors have diagnosed, after writing the report, AI can check all the reports, compare the corresponding images, find a doctor Problems that may be overlooked. If problems are found, the AI ​​can alert the doctor so that the doctor can check again the next day."

“A new AI application is used for child growth assessment. The AI ​​system can estimate the bone age of a child in one second. Of course, more common is the intelligent detection of fractures. AI can locate the position of the ribs. Separate the ribs and label them and generate a structured report."

"In addition to the application just described, AI can delineate non-small cell lung cancer in the lungs, and only 0.3 seconds can be used to outline non-small cell lung cancer, which can save a lot of time for imaging studies. In addition, through CT scan of the chest, AI can not only do lung nodule detection, but also can make cardiac hypertrophy warning, even fracture early warning."

Applications need platform integration

In his speech, Professor Shen said that Lian Ying Intelligence is currently building a medical image depth research platform that includes independent modules such as image segmentation, target detection, image classification, image registration, and image mapping. Each module will be standardized. , but also free to combine. In this way, when this platform is built, the long process from algorithm development to application deployment will become simple, and the deployment of AI products will become faster and faster.

Food flavoring

Tartaric Acid, Vanillin, Menthol, Menthol Crystal, Flavoring agent, Baking powder

Xi'an Gawen Biotechnology Co., Ltd , https://www.ahualyn-bios.com