"Insufficient inhalation", "non-iatrogenic foreign body", "poor exposure"... This is the "quick evaluation" of the X-ray film taken by medical imaging technicians by artificial intelligence (AI) systems. With the AI's "real-time full inspection", technicians can immediately discover problems and correct them, eliminating the trouble of asking patients to "go to the hospital again" to re-shoot the film. is a new application scenario of the "combined 5G medical alliance image collaborative innovation platform" led by Zhongshan Hospital Affiliated to Fudan University and jointly built by Shanghai Lianying, Lianying Intelligent, Shanghai Unicom and other units. This also means that the first medical imaging AI application across regions, institutions and categories is implemented in China.
is of revolutionary significance to the doctor side. "In the past, we read films manually. As the reading time increases, the efficiency of reading films will decrease. But AI will not be tired. It reads the films 24 hours a day and uses lung nodules, coronary plaques and other various types of pulmonary nodules, coronary artery plaques, etc. The lesions are marked and there is almost no missed. In this way, the doctor will review them again, focusing on suspicious lesions, and the efficiency and accuracy of the film reading are greatly improved." Zeng Mengsu, director of the Radiology Diagnosis Department of Zhongshan Hospital, said.
At present, the platform has high-quality labeled data sets of 100,000, AI real-time quality control, 170,000 AI-assisted diagnosis, 220,000 AI-assisted diagnosis, 15% of diagnostic sensitivity, and 30% of video reading efficiency.
Just as a doctor's ability will be upgraded, this "AI doctor" who assists in reading the film will also be upgraded, but the "learning speed and ability" may be several times that of a human doctor.
"With big data, collaborative annotation and other functions, we can quickly obtain clinical data in the real world, and then we will make our scientific conjectures into 'model training' through AI, and then go to the real world to verify, and the successful model becomes The latest digital tools can be applied to clinical practice to achieve the upgrade of this AI doctor." Zeng Mengsu said.
integrated 5G medical alliance image collaborative innovation platform has also opened up clinical applications and scientific research, helping doctors achieve cross-campus and cross-level collaboration, and achieving efficient data annotation and scientific research exploration. It has formed clinical applications to inspire scientific research and scientific research for clinical applications to provide clinical applications. Feedback and thus improve application performance collaborative mode.
According to reports, the platform comprehensively uses new generation technologies such as artificial intelligence, cloud computing, big data, 5G mobile communications, and has created the first one-stop, full-link medical imaging AI across regions, institutions, and categories in China. , standardized application solutions. , Qian Kun, director of the Planning and Management Center of the Information and Intelligent Development Department of Zhongshan Hospital, introduced that the platform integrates AI data governance, AI application governance, and AI capability opening, and uses the AI collaborative innovation of image AI within the scope of Fudan Zhongshan Medical Alliance as a scenario to build The industrial innovation cycle of technological improvement, data accumulation, knowledge accumulation, and application sharing will promote the application and implementation of imaging AI in medical institutions at all levels.
"Medical + AI" opens up more possibilities. These "AI doctors" that embody high-level clinical knowledge of physicians can be replicated infinitely, accessed through the platform, and then moved to primary medical institutions. With the help of the auxiliary reading of "AI doctors", the diagnosis and treatment level and efficiency of grassroots medical personnel will be greatly improved, truly realizing the low-cost expansion of high-quality diagnosis and treatment resources, and achieving extensive radiation and deep sinking of resources.
reporter learned from Zhongshan Hospital that the project was put into trial operation in 2020 and officially passed the project acceptance of the Shanghai Economic and Information Commission in 2021. The results of the project have been applied to many medical institutions in the Yangtze River Delta, Xiamen, Fujian, Kashgar, Xinjiang and other regions, with the cumulative number of cases on the platform exceeding 3 million.
Not long ago, the "Medical Alliance Image Collaborative Innovation Platform with 5G" was selected into the TOP30 list of the 2022 SAIL Award (Leader of Excellent Artificial Intelligence). Zhongshan Hospital is the only comprehensive hospital selected for the list. The award is the highest award at the World Artificial Intelligence Conference. At present, the project has obtained 5 invention patents, was selected as the National Health Commission's "5G+ Medical and Health Application Pilot Project" project, and won three awards including the "2020 Typical Enterprise Cloud Cases" of the Ministry of Industry and Information Technology.
Gu Jianying, deputy director of Zhongshan Hospital, said, "Zhongshan Hospital will use this as a basis to continue to be patient-centered and disease-centered, promote technological innovation, and carry out meta-image of "shadow brain" 5G+ smart medical ecosystem" ' and other projects continue to promote the application of medical imaging artificial intelligence technology in the health field, and drive the overall improvement of medical level in the entire industry."
Author: Tang Wenjia Feng Ying
Editor: Tang Wenjia
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-