Artificial Intelligence and the Origin of Medicine
1956, John? John McCarthy, Marvin? Marvin Liminsky, Claude? Elwood. Claude elwood shannon, Nathaniel? Four scholars in Rochester initiated an AI summer seminar, pointing out that the research goal of "AI" is to realize a machine that can simulate human beings. Machines can use language, have the ability to understand, can complete tasks that only humans can complete, and constantly improve the machine itself. This concept has been developed for 65 years.
"Seeing the definition of AI, it is not difficult to find that AI is closely related to brain science and clinical neuroscience. The design of AI network is also imitating the human brain. The ultimate AI is to realize brain like computing. " Wang Yongjun said.
When AI first appeared, it had been applied to the medical field, and the earliest application mode was electronic medical record. After 20 10 years, AI will be used to process genomic big data and realize natural language recognition on electronic medical records.
Wang Yongjun introduced that since 20 19, AI technology has been applied in the medical field by looking for new drug targets and drug research and development.
In 2020, Nature Medicine published two important guidelines, SPIRIT-AI EXTENSION and CONSORT-AI EXTENSION, which provided a foundation for AI to enter clinical application and were two important programmatic documents to promote AI clinical application.
AI makes stroke patients no longer limited to the golden period of rescue.
AI is a good helper for doctors, because its appearance has changed the rescue mode of stroke patients.
As we all know, when a stroke occurs, it is necessary to carry out rescue at prime time-intravenous thrombolysis within 4.5 hours and mechanical embolectomy within 6 hours. Once this time is exceeded, it is easy to cause disability or even death.
Wang Yongjun said: "After stroke, there is a core necrosis area in the middle of middle cerebral artery occlusion, and there is a circle of ischemia outside, but there is no necrosis. This is the window period for doctors to treat. Under normal circumstances, every 65,438+000 grams of brain tissue flows through about 65,438+000 ml of blood every minute. When it is reduced to 20 ml, the internal and external exchange of sodium and potassium ions in cell membrane will stop. Below 65,438+00 ml, the cell membrane will start to rupture. We call the blood flow period of 10 to 20 ml a semi-dark area, which provides a window for doctors to rescue. "
As early as 1980s, it was a difficult problem to recognize penumbra. Without intelligent means, doctors can only recognize penumbra by watching films, which affects the speed of first aid. With AI technology, this traditional way has been completely changed, which not only can better identify the penumbra, but also does not need to consider the rescue time window.
From knowledge-driven to data-driven drug research and development
Drug research and development is a scientific research with very high difficulty coefficient. It takes more than ten years or even decades for each new drug to go from research and development to market application, and the emergence of AI has accelerated this process.
"Before 20 15, our drug research and development model was a traditional model. The general process is to determine the disease target first, then carry out animal cell model, organ model and animal model experiments, and then carry out clinical trials. This process takes more than ten years. After 20 15, AI began to be applied to drug research and development, subverting the whole R&D process. We interact image data, histological data, clinical data and environmental data to form big data, and use AI to find the mathematical relationship between these data, thus making innovative drugs and innovative reagents. "
The transformation of this process also shows that the drug research and development model has changed from knowledge-driven to data-driven. Wang Yongjun said: "The reverse drug research and development model of clinical big data and multi-omics, abbreviated as BBB model, combines knowledge-driven and data-driven, which improves the performance of finished drugs by 100 times and shortens the research and development time by half. This is the new change that AI has brought us. "
In addition, AI is also used in many aspects of the medical field, such as the use of Da Vinci surgical robots, as well as Japanese AI escort equipment and ward-round robots, which improve efficiency and reduce potential harm to doctors through machine work.
Wang Yongjun said: "Artificial intelligence is developing rapidly in the field of clinical neuroscience. One of the reasons is that we have made AI a partner, which enables clinicians to use AI technology earlier and more smoothly, and also brings benefits to patients. "
Typesetting: Sorry.