Asan Medical Center at the Forefront in Critical Illness Care with AI, Big Data - 네이트

Advances in artificial intelligence (AI) and big data are rapidly reshaping modern medicine, shifting the paradigm toward data driven precision care. While early medical AI applications largely assisted physicians with image interpretation, next generation systems are increasingly capable of analyzing vast clinical datasets in real time to predict patient deterioration, guide surgical strategies, and improve safety in high-risk medical situations. Asan Medical Center, one of South Korea’s leading institutions for treating severe and complex diseases, is accelerating its transition toward a smart hospital model to keep pace with this transformation. The hospital recently established a dedicated Office for AI Innovation Support and has significantly expanded its research and commercialization efforts in data-driven medicine. Over the past three years, the hospital has recorded 34 technology transfer deals and filed 77 patents in the fields of AI and big data. These achievements are laying the technological groundwork for more precise and safer personalized treatments for critically ill patients. ■ Converting complex lesions into data to maximize diagnostic accuracy with AI Asan Medical Center is applying AI technology to visualize subtle lesions and blood flow changes that are difficult to detect with the naked eye, translating them into quantitative data that improves diagnostic precision. A research team led by Professors Dong Hyun Yang and Hyun Jung Koo of the Department of Radiology and June-Goo Lee of the Department of Convergence Medicine has developed an AI based quantitative analysis model for the aorta. The system automatically segments the aorta in three dimensions using chest CT images and precisely measures the diameter, area, and volume of each segment. The model incorporates the latest anatomical standards based on the SVS/STS classification system, enabling detailed segmentation of the aorta and providing reference values for each region. This allows personalized comparisons tailored to individual patient characteristics such as age and sex. Once fully implemented in clinical practice, the technology could move diagnostics beyond the traditional method of identifying abnormalities based on a single threshold value. Instead, physicians can perform individualized analysis based on specific segments and patient characteristics, potentially improving early detection rates of serious conditions such as aortic dissection and aneurysms. The system also reduces variability among observers during manual measurements, improving consistency in image interpretation. It is expected to play a key role in determining optimal treatment timing by enabling objective long term monitoring of patient outcomes following surgery or intervention. The technology was designated an innovative medical technology by the Ministry of Health and Welfare in June 2024 and is already being applied in clinical settings. Its use is expected to expand to diagnostic support and long term monitoring systems. ■ AI safety network protects the golden hour for critically ill patients Asan Medical Center is also building an intelligent safety network that leverages AI’s analytical power to assist physicians in making faster and more accurate decisions during emergencies where every second counts. A team led by Professor Sung-Hoon Kim of the Department of Anesthesiology and Pain Medicine has developed a diagnostic support technology that predicts neurological outcomes after cardiac arrest. The system analyzes real-time electroencephalogram (EEG) signals from comatose patients using deep learning and quantifies the probability of neurological recovery. Even in situations where continuous interpretation by specialists is difficult—such as overnight shifts or emergency conditions—the AI system can monitor EEG patterns around the clock, significantly improving the objectivity of prognosis assessments. The hospital is also developing a deep learning based arrhythmia detection system capable of automatically identifying abnormal heart rhythms and evaluating their risk levels in real time in environments such as operating rooms, emergency departments and intensive care units. The system is designed to operate reliably even in noisy clinical environments where electrosurgical devices or patient movement may interfere with signals. Clinical validation has already been completed using ECG data from the hospital’s operating rooms and intensive care units, and follow-up studies are underway to analyze risk factors and treatment outcomes more comprehensively. Another AI technology developed by Professor Dong-Wha Kang of the Department of Neurology predicts the onset time of ischemic stroke, offering a critical solution for preserving the golden treatment window for emergency stroke patients. Because clot-dissolving therapy must typically be administered within 4.5 hours of stroke onset, many patients miss treatment opportunities when the exact onset ti