한빛사논문
- 조회1,113
Seung Wook Lee1, Hyung-Chul Lee2, Jungyo Suh3, Kyung Hyun Lee4, Heonyi Lee5, Suryang Seo6, Tae Kyong Kim7, Sang-Wook Lee8,* and Yi-Jun Kim9,*
1School of Medicine, Kyungpook National University, Daegu, Republic of Korea. 2Department of Anesthesiology and Pain Medicine, Seoul National University Hospital, Seoul National University College of Medicine, Seoul, Republic of Korea. 3Department of Urology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea. 4Department of Digital Health, SAIHST, Sungkyunkwan University, Seoul, Republic of Korea. 5Bioinformatics Collaboration Unit, Department of Biomedical Systems informatics, Yonsei University College of medicine, Seoul, Republic of Korea. 6Department of Nursing, SMG-SNU Boramae Medical Center, Seoul National University College of Medicine, Seoul, South Korea. 7Department of Anesthesiology and Pain Medicine, SMG-SNU Boramae Medical Center, Seoul National University College of Medicine, Seoul, South Korea. 8Department of Anesthesiology and Pain Medicine, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea. 9Institute of Convergence Medicine, Ewha Womans University Mokdong Hospital, Seoul, Republic of Korea.
*Corresponding author.
Abstract
Accurate prediction of postoperative mortality is important for not only successful postoperative patient care but also for information-based shared decision-making with patients and efficient allocation of medical resources. This study aimed to create a machine-learning prediction model for 30-day mortality after a non-cardiac surgery that adapts to the manageable amount of clinical information as input features and is validated against multi-centered rather than single-centered data. Data were collected from 454,404 patients over 18 years of age who underwent non-cardiac surgeries from four independent institutions. We performed a retrospective analysis of the retrieved data. Only 12–18 clinical variables were used for model training. Logistic regression, random forest classifier, extreme gradient boosting (XGBoost), and deep neural network methods were applied to compare the prediction performances. To reduce overfitting and create a robust model, bootstrapping and grid search with tenfold cross-validation were performed. The XGBoost method in Seoul National University Hospital (SNUH) data delivers the best performance in terms of the area under receiver operating characteristic curve (AUROC) (0.9376) and the area under the precision-recall curve (0.1593). The predictive performance was the best when the SNUH model was validated with Ewha Womans University Medical Center data (AUROC, 0.941). Preoperative albumin, prothrombin time, and age were the most important features in the model for each hospital. It is possible to create a robust artificial intelligence prediction model applicable to multiple institutions through a light predictive model using only minimal preoperative information that can be automatically extracted from each hospital.
논문정보
- Research article
- 2022년 07월 (BRIC 등록일 2022-07-14)
- 국내 연구진
- 의약학 > 응용의학
관련 웨비나
![수술 후 사망률 예측 인공지능 모델의 다기관 검증연구 [npj Digit. Med.]](/upload/board/files/onseminar/thumb/thumb_0000432.png)
학술웨비나
수술 후 사망률 예측 인공지능 모델의 다기관 검증연구 [npj Digit. Med.] 최근 인공지능 기술의 발달로 인하여 의료 분야에서도 이를 활용하고자 하는 연구들이 활발하게 이루어지고 있습니다. 하지만 인공지능 기술의 대표적인 한계점은 overfitting이라는 문제로 인하여 다른 임상 환경에서 적용하기 어려운 문제들이 있었습니다. 본 연구는 국내 4개의 대규모 데이터를 기반으로 서로 다른 예측 모델들로 서로 교차 검증함으로써 인공지능 모델의 다기관 검증에 관한 매우 중요한 이정표를 제시한 연구입니다. 특히, 이번 연구에서는 EMR에서 자동으로 추출이 가능한 혈액검사 결과나 단순한 demographic 정보들을 활용하여 수술 후 30일 이내 사망률을 예측하는데 있어서 적은 수의 입력 변수들만 가지고도 많은 수의 데이터 변수들을 이용하는 앞선 연구들의 예측 모델들에 비해서 결코 예측 성능이 떨어지지 않았습니다. 무엇보다 다기관 검증을 통하여 오히려 적은 수의 변수를 가지고 모델링하는 것이 좀 더 robust한 인공지능 예측 모델 개발이 가능함을 보여주고 있습니다.- 이상욱(울산대학교 의과대학, 서울아산병원)
댓글 작성 로그인
팝업 닫기관련 링크
연구자 키워드
관련분야 논문보기
해당논문 저자보기
