Prof. Bin SHENG and his team published a review article in Cell Reports Medicine, entitled “Artificial intelligence in diabetes management: Advancements, opportunities, and challenge”.
[About Cell Reports Medicine]
Cell Reports Medicine (IF=14.3) is a premium journal from Cell Press publishing cutting-edge research in translational and clinical biomedical sciences that inform and influence human health and medicine. The content published in Cell Reports Medicine reaches a broad range of scientists and clinicians across the spectrum of medical disciplines, ensuring that the research work will be both visible and accessible. The journal publishes original research that ranges from exciting concepts in human biology, health, and disease to all phases of clinical work.
[Research content]
Graphic Abstract
The increasing prevalence of diabetes has become a global public health concern in the 21st century. Approximately 600 million people had diabetes globally, and China accounts for the largest number of people with diabetes (140 million). The increasing prevalence of diabetes, high avoidable morbidity, and mortality due to diabetes and diabetic complications, and related substantial economic burden make diabetes a significant health challenge. Thus, prevention, early diagnosis and early treatment are of great importance in diabetes care. Recent advancements in digital health technologies (DHTs), especially artificial intelligence (AI), provide a significant opportunity to achieve better efficiency in diabetes care, which may diminish the increase in diabetes-related healthcare expenditures. Based on advanced technologies of machine learning (ML), computer vision, and virtual reality, the new data-driven model is a promising tool that can transform diabetes care and improve the lives of millions of people worldwide.
The study reviewed the recent progress in the application of AI in the management of diabetes and then discussed the opportunities and challenges of AI application in clinical practice. Furthermore, the study explored the possibility of combining and expanding upon existing digital health technologies to develop an AI-assisted digital healthcare ecosystem that includes the prevention and management of diabetes.
Figure 1 Current applications of AI in multiple areas in diabetes management
AI is a broad branch of computer science that develops theories, methods, technologies, and application systems to simulate, extend, and expand human intelligence in machines. Machine learning (ML) is a subcategory of AI that uses statistical techniques to build intelligent systems.Deep learning (DL), using advanced ML techniques, has achieved significant success in computer vision and natural language processing tasks, primarily attributed to its excellent feature extraction and pattern recognition capabilities.
The application of AI in diabetes care and research has been widely explored in basic biomedical research, translational research, and clinical practice (Figure 1). For instance, AI technology can help predict the onsets of diabetes, screen and classify diabetes, provide comprehensive health education, and offer recommendations on diet, physical, and drug therapy. Likewise, AI plays a significant role in prediction, screening, and management of diabetic complications. For example, based on the world's largest-scale community-based database of fundus images, the construction of which was leaded by Academician Weiping Jia, Professor Bin Sheng and his research team in Shanghai Jiao Tong University developed a DeepDR-LLM system. This system can provide diabetic retinopathy (DR) screening and patient disposition management recommendations, based on retinal photographs and clinical metadata. The research achievement “A deep learning system for detecting diabetic retinopathy across the disease spectrum” was published in Nature Communications, sub-journal of Nature. It was selected as an Editors' Highlight, and recognized as a Highly Cited Paper in the Essential Science Indicators (ESI). The work was acclaimed as “one of the most significant technological innovations and research achievements in the field of translational and clinical research recently”..
Despite the great potential of AI technology revealed in diabetes care, its application in clinical practice still faces many obstacles: 1) poor quality of data; 2) poor technology design; 3) lack of clinical integration; 4) privacy concern; 5) non-adherence; 6) imperfection of laws and regulations, etc.
To address these challenges of AI application, the following solutions are warranted: 1) ensure the data quality in training AI algorithms; 2) prioritize user experience and engage in iterative design; 3)align with clinical practice to provide meaningful decision support; 4) enhance security design to protect patient privacy; 5) explore ways to improve adherence; 6) strengthen legal supervision for AI healthcare (Figure 2).
Figure 2 Opportunities, Chanllenges and Future Directions of AI application in diabetes management
At present, AI technology has achieved a good clinical transformation in DR screening and blood glucose monitoring. For instance, the first well-established device is IDx-DR, approved by the FDA in 2018 for its high diagnostic performance in clinical trials. Another example is the Guardian Connect System (manufactured by Medtronic). This system is characterized by using AI to predict a hypoglycemic attack an hour in advance based on the CGM data and alerts the patient. From the above cases, it can be seen that AI system can assist healthcare professionals by providing timely and precise supplementary decision so as to improve health management of patients.
Integrating AI-based DHTs will probably become increasingly feasible in the future as technology improves, and such integration will enable new models of diabetes care. Here, the study proposed the construction of an AI-assisted digital healthcare ecosystem for diabetes management (Figure 3). This ecosystem consists of several essential sessions enabled by AI: 1) Recognize the risk factors of diabetes and predict the risk of diabetes onset in the general public. Based on the risk and modifiable risk factors, the system can provide personalized suggestions and continuous monitoring to control these contributing factors; 2) Implement diabetes screening in high-risk populations; 3) Assist medical practitioners and patients in the basic management of diabetes: health education, medical nutrition therapy, physical therapy, and drug therapy; 4) Provide the prediction, screening, and management of diabetic complications.
Figure 3 AI-assisted digital healthcare ecosystem for diabetes management
In conclusion, AI has the potential to optimize diabetes care by providing personalized, precise, and data-driven support to patients and healthcare professionals. By addressing the challenges and capitalizing on the opportunities, AI could play a pivotal role in transforming diabetes care and improving the lives of millions of people worldwide. This research review is of reference value for advancing the application of AI technology in diabetes management in the future.
Reference:
Zhouyu Guan, Huating Li, Ruhan Liu, Chun Cai, Yuexing Liu, Jiajia Li, Xiangning Wang, Shan Huang, Liang Wu, Dan Liu, Shujie Yu, Zheyuan Wang, Jia Shu, Xuhong Hou, Xiaokang Yang, Weiping Jia, Bin Sheng, Artificial intelligence in diabetes management: Advancements, opportunities, and challenges, Cell Reports Medicine, 2023, https://doi.org/10.1016/j.xcrm.2023.101213.
Co-first Authors:
Zhouyu GUAN, MD student in endocrinology and metabolism at Shanghai Jiao Tong University School of Medicine. His research interests include the artificial intelligence management, deep learning, and digital healthcare applications in diabetes and diabetic complications. He has published academic papers as a first author (including co-first author) in Cell Reports Medicine andFrontiers in Public Health.
Huating LI, Professor in Shanghai Sixth People’s Hospital Affiliated to Shanghai Jiao Tong University School of Medicine. His research interests include endocrinology and metabolism. She has published over 50 SCI papers, including in journals including Cell Metabolism, Science Translational Medicine, and Nature Communications.
Ruhan LIU, PhD in Computer Science and Technology from Shanghai Jiao Tong University, supervised by Professor Bin Sheng. And she is currently an Assistant Professor at Furong Laboratory, China. Her main research interests include medical artificial intelligence. She has published 8 articles in journals including Nature Communications, IEEE Trans. Medical Imaging, IEEE Trans. Biomedical Engineering, Cell Reports Medicine, and Patterns.
Co-last Authors:
Bin SHENG, Full Professor in Department of Computer Science and Engineering, School of Electronics Information and Electrical Engineering (SEIEE) , Shanghai Jiao Tong University. He serves as the co-chair of AI Challenges in ISBI 2020, MICCAI 2022, and MICCAI 2023. He is also the Managing Editor of The Visual Computer, and Associate Editor of IEEE Trans. CSVT, The Visual Computer, and Virtual Reality and Intelligent Hardware (VRIH). Prof. Sheng has published more than a hundred technical papers refereed in journals including Nature Communications、IEEE TPAMI、IEEE TVCG, IEEE TIP, IEEE TMI、Medical Image Analysis, and in international conferences such as IEEE Virtual Reality、ICCV、AAAI and ACM Multimedia.
Weiping JIA, Academician of Chinese Academy of Engineering and Chair Professor of Shanghai Jiao Tong University. She is the Editor-in-Chief of the Chinese Journal of Internal Medicine. Her research interests include precise diagnosis and treatment of diabetes. She has published over 300 papers in journals including BMJ, Diabetes Care, and Lancet Diabetes Endocrinol.