Cardio-cerebrovascular Disease Predicted by Artificial Intelligence After Threatening Surgery for Elderly Patients
Apr 29, 2025
With the recent aging and development of medical technology, the number of patients over 65 years of age undergoing various surgeries such as cancer surgery and joint surgery has increased significantly, and an artificial intelligence model has been developed by domestic researchers that can predict cardio-cerebrovascular disease, a major postoperative complication, in advance in elderly patients.
A research team led by Professor Seo Jeong-won of the Department of Circulatory Medicine at Seoul National University Bundang Hospital (Professor Kwon Joo-sung of the Department of Circulatory Medicine, Professor Ahn Hyung-beom, and Professor Yoo Soo-young of the Digital Healthcare Research Project) said they have developed a machine learning-based algorithm to predict the risk of cardiovascular disease after surgery by analyzing medical records of elderly patients undergoing surgery except for heart surgery.
Cardio-cerebrovascular diseases such as myocardial infarction and stroke are one of the postoperative complications that cannot be overlooked in elderly patients.
The older you are, the more chronic diseases such as high blood pressure, diabetes, and heart disease are common, and exposure to general anesthesia, bleeding during surgery, and inflammatory reactions puts a great burden on the cardio-cerebrovascular relationship.
Until now, medical sites have evaluated the risk of cardio-cerebrovascular disease in patients using a tool called 'Revised Cardiac Risk Index (RCRI)' before surgery.
However, RCRI is pointed out as a limitation in that it evaluates using only limited information such as age, history of heart disease, and type of surgery. In particular, important information such as blood test results, drugs being taken, and past diagnosis names are missing, which makes it difficult for medical staff to accurately evaluate the actual patient's risk.
To overcome these limitations, the research team developed a model that precisely predicts cardio-cerebrovascular complications that can occur within 30 days of general surgery except for heart surgery by analyzing comprehensive information such as blood test results, underlying diseases, drugs taken, and surgery types recorded in patients' electronic medical records (EMRs) through artificial intelligence.
Data from 46,000 patients at Seoul National University Bundang Hospital were used for the study, and external verification was performed through a cohort of Asan Medical Center in Seoul.
The model developed by the research team has a predictive accuracy (AUROC, curved area) of up to 0.897, far superior predictive power compared to RCRI (0.704), the existing standard evaluation tool. These results are meaningful in that they can quickly and simply predict cardio-cerebrovascular diseases after surgery in the field without a separate detailed examination, and they are expected to be expanded to various hospitals as they were developed through a standardization process.
Professor Seo Jeong-won said, `Even at a similar age, elderly patients have large deviations in their health conditions, so if they can accurately predict the risk of cardio-cerebrovascular complications after surgery, it is very helpful for patient safety"We plan to develop the model in a way that medical staff can use easily and quickly in connection with the hospital's system," he said.
Meanwhile, the study was conducted with the support of the Korea Research Foundation under the Ministry of Science and ICT, and the results of the study were published in the prestigious international academic journal 『Journal of Medical Internet Research" in the field of digital healthcare.
A research team led by Professor Seo Jeong-won of the Department of Circulatory Medicine at Seoul National University Bundang Hospital (Professor Kwon Joo-sung of the Department of Circulatory Medicine, Professor Ahn Hyung-beom, and Professor Yoo Soo-young of the Digital Healthcare Research Project) said they have developed a machine learning-based algorithm to predict the risk of cardiovascular disease after surgery by analyzing medical records of elderly patients undergoing surgery except for heart surgery.
Cardio-cerebrovascular diseases such as myocardial infarction and stroke are one of the postoperative complications that cannot be overlooked in elderly patients.
The older you are, the more chronic diseases such as high blood pressure, diabetes, and heart disease are common, and exposure to general anesthesia, bleeding during surgery, and inflammatory reactions puts a great burden on the cardio-cerebrovascular relationship.
Until now, medical sites have evaluated the risk of cardio-cerebrovascular disease in patients using a tool called 'Revised Cardiac Risk Index (RCRI)' before surgery.
However, RCRI is pointed out as a limitation in that it evaluates using only limited information such as age, history of heart disease, and type of surgery. In particular, important information such as blood test results, drugs being taken, and past diagnosis names are missing, which makes it difficult for medical staff to accurately evaluate the actual patient's risk.
To overcome these limitations, the research team developed a model that precisely predicts cardio-cerebrovascular complications that can occur within 30 days of general surgery except for heart surgery by analyzing comprehensive information such as blood test results, underlying diseases, drugs taken, and surgery types recorded in patients' electronic medical records (EMRs) through artificial intelligence.
Data from 46,000 patients at Seoul National University Bundang Hospital were used for the study, and external verification was performed through a cohort of Asan Medical Center in Seoul.
The model developed by the research team has a predictive accuracy (AUROC, curved area) of up to 0.897, far superior predictive power compared to RCRI (0.704), the existing standard evaluation tool. These results are meaningful in that they can quickly and simply predict cardio-cerebrovascular diseases after surgery in the field without a separate detailed examination, and they are expected to be expanded to various hospitals as they were developed through a standardization process.
Professor Seo Jeong-won said, `Even at a similar age, elderly patients have large deviations in their health conditions, so if they can accurately predict the risk of cardio-cerebrovascular complications after surgery, it is very helpful for patient safety"We plan to develop the model in a way that medical staff can use easily and quickly in connection with the hospital's system," he said.
Meanwhile, the study was conducted with the support of the Korea Research Foundation under the Ministry of Science and ICT, and the results of the study were published in the prestigious international academic journal 『Journal of Medical Internet Research" in the field of digital healthcare.
|
This article was translated by Naver AI translator.
