Medical care transformed by artificial intelligence

At present, a new round of scientific and technological revolution and industrial transformation led by artificial intelligence is in the ascendant, and life sciences and medical health are one of the most important application places in the field of artificial intelligence. On April 28, the parallel forum of the 2024 Zhongguancun Forum Annual Conference "Reshaping Healthcare: Innovative Artificial Intelligence Transformation in Medicine" was held. The panelists believed that AI-empowered clinical research, new drug research and development, and medical device innovation can greatly improve disease prevention, medical diagnosis, and treatment, but how to meet clinical needs through AI algorithms, including use standards, privacy, and transparency, still needs to be explored.

Beijing ranks first in the country in the number of registered medical AI products

"18 companies have been approved for 27 types of registration certificates for medical AI-assisted diagnosis scenarios, accounting for 34% of the national total," Deng Pingji, deputy director of the Beijing Municipal Health Commission, shared such a set of data at the parallel forum on "Reshaping Healthcare: Innovative AI Transformation in Medicine".

At present, Beijing leads the country in the development level of medical artificial intelligence, and the number of medical artificial intelligence enterprises and registered products ranks first in the country. According to Deng Pingji, many medical institutions, physicians and scientists are deeply involved in the research and development of artificial intelligence products, and the application of artificial intelligence medical products involves different fields and aspects such as medical auxiliary decision-making, health management, patient intelligent service, medical robots, hospital management, precision medicine, etc., especially in the analysis and application of medical imaging, medical documents, audio and video, and multi-omics data analysis.

Deng Pingji said that driven by new theories and new technologies, artificial intelligence has presented new characteristics such as self-learning, cross-border integration, human-machine collaboration, and group intelligence openness, which is having a significant and far-reaching impact on economic development, social progress, and global governance. Life sciences and medical health are one of the most important application sites in the field of artificial intelligence, AI empowers clinical research, new drug research and development, and medical device innovation, which can greatly improve the level of disease prevention, medical diagnosis and treatment, improve the accessibility and inclusiveness of health services, improve service quality, reduce service costs, and its development prospects are very broad and the potential is unlimited.

The Beijing Municipal Health Commission will study and formulate policies and measures that meet the needs in accordance with the deployment of the Municipal Party Committee and the Municipal Government, further promote the research and development and application of medical artificial intelligence products, strengthen policy support, strengthen policy guarantees, and strengthen governance, so as to better promote the scientific and technological innovation and transformation of medical achievements and serve the health and industrial development of the people in the capital.

Shorten the R&D cycle of new drugs and reduce costs

An important role of artificial intelligence in the medical industry is to shorten the research and development cycle of new drugs and reduce costs. Dr. Lei Tang, Head of Sanofi Translational Medicine China, shared how the company is using AI to improve the quality and speed of small molecule drug discovery.

As we all know, drug R&D is a very long process from drug target determination, lead compound screening, lead compound optimization, clinical trials, registration and marketing. In general, it takes about 10 years and a lot of money to develop a new treatment, Tang said. According to publicly available data, pharmaceutical companies spend an average of $2.6 billion on this process.

Sanofi is trying to enhance R&D productivity with the help of AI capabilities by deploying digital technologies. The company uses digital and AI technologies in translational medicine, clinical research, and regulatory and patient safety phases. In terms of small molecule drug discovery, scientists have developed the targeted immunity engine TIE, which analyzes and organizes data, including deep mining of data from various sources, and even data from competing drugs. The company's research on ulcerative colitis has been used to predict more than 50 targets, and more than 50 target hypotheses have been generated in less than 12 months, with seven innovative targets advanced into the research pipeline, which is a result achieved in 2023 alone.

Tang Lei believes that AI models can improve research efficiency. At present, AI models are used in the screening and exploration research of more than eighty percent of small molecule compounds, which greatly accelerates the speed of research, and if they can be brought into clinical research quickly, they can benefit patients faster.

Professor Zhang Zongjiu, executive vice president of the Institute of Hospital Management of Tsinghua University, said that the application of AI in more scenarios has broken geographical restrictions, made high-quality medical resources benefit every corner, shortened the research and development cycle of new drugs, and reduced costs.

Rapid medical imaging diagnostic results

Huang Feng, Chief Digital Officer of GE Healthcare China, mentioned the application of AI in medical imaging during the roundtable sharing session. The first successful and innovative application of AI in the medical field is in the field of imaging. According to Huang, the industry has spent a long time exploring the accuracy of images. However, with the advent of AI, diagnostic results can be obtained quickly and accurately.

Huang Feng believes that China has a very good ecosystem and many excellent AI start-ups, and GE Healthcare has also cooperated with these companies on some products, which are developed in China and adopted globally. According to a study, 48% of patients worldwide can accept AI, and in China, this acceptance rate exceeds 70%.

Looking to the future, GE Healthcare is also experimenting with large-scale language models and working with engineers to help customers. However, Huang Feng also mentioned that because the medical field is very complex and there are strict norms to be followed in serious fields, even if there are so many models, it is difficult for customers to quickly understand various AI applications by themselves, and engineers have limitations in their understanding.

Dr. Noa Dagan, director of AI medicine at the Clarit Institute and associate professor at Ben-Gurion University, shared the success of AI-driven medicine, according to her, because the hepatitis C screening rate in high-risk groups is not high, researchers have developed AI algorithms that can help predict hepatitis C-positive patients, thereby improving the compliance and efficiency of screening. According to Professor Wang Yilong, vice president of Beijing Tiantan Hospital affiliated to Capital Medical University, AI is playing an increasingly important role in the prevention, diagnosis, treatment and management of neurological diseases. In terms of prevention, AI can predict the occurrence and severity of Parkinson's disease at an early stage by analyzing big data. In terms of diagnosis, AI technology can quickly analyze imaging data, predict the pathogenesis of cerebrovascular diseases, and assist in individualized and precise treatment.

Privacy, transparency, and data governance remain priorities

AI has enormous potential, but in this particular field of healthcare, there is a need to balance its potential and responsibility. Professor Xiao Ruiping, associate editor-in-chief of NEJM and dean of the School of Future Technologies at Peking University, said that medical AI has the potential to improve treatment outcomes, improve medical information exchange, and improve the efficiency of healthcare systems, but AI-related ethical issues also need to be considered, such as privacy, transparency, and data sharing. She called for the development of standards for the use of AI in healthcare and for transparency.

In the view of Professor Cai Xiujun, President of Sir Run Run Shaw Hospital affiliated to Zhejiang University School of Medicine, "artificial intelligence + medical care" needs some technological innovation policies. He emphasized the importance of three aspects: the cultivation of interdisciplinary talents between artificial intelligence and medicine is very important, the convenience of payment is very important, and the standardization of data and information security are very important.

Cai Xiujun believes that AI is a very deep discipline, and clinical practice is also a very deep discipline. Even if there is computing power, it will be a big problem if you don't know what clinical needs are. Therefore, the solution to the problem is mutual communication between the two sides. He suggested that colleges and universities should cultivate some interdisciplinary talents and give full play to the role of hospital IT teams.

From the perspective of clinicians, Wang Yilong pointed out that it is very important to provide multimodal high-quality data in the future, and clinicians should put forward the real demand for AI, encourage the cooperation of more disciplines, and hope that AI will help medical care in the whole chain and all dimensions.

Dr. Charlotte Haug, Managing Editor-in-Chief of NEJM AI, emphasized the importance of AI regulation. How to ensure global regulatory consistency is a question that needs to be considered, otherwise an AI product can only be used in the research and development site. It is also important to ensure that data is used appropriately and that effective data is shared responsibly. In healthcare, the data used is sensitive patient data, and it's not easy to share it properly. Patient data and information privacy is the first issue to be addressed in the large-scale use of AI in medicine.