Accurate prediction of postoperative complications can inform shared decisions regarding prognosis, preoperative risk-reduction, and postoperative resource use. We hypothesized that multi-task deep learning models would outperform random forest models in predicting postoperative complications, and that integrating high-resolution intraoperative physiological time series would result in more granular and personalized health representations that would improve prognostication compared to preoperative predictions. In a longitudinal cohort study of 56,242 patients undergoing 67,481 inpatient surgical procedures at a university medical center, we compared deep learning models with random forests for predicting nine common postoperative complications using preoperative, intraoperative, and perioperative patient data. Our study indicated several significant results across experimental settings that suggest the utility of deep learning for capturing more precise representations of patient health for augmented surgical decision support. Multi-task learning improved efficiency by reducing computational resources without compromising predictive performance. Integrated gradients interpretability mechanisms identified potentially modifiable risk factors for each complication. Monte Carlo dropout methods provided a quantitative measure of prediction uncertainty that has the potential to enhance clinical trust. Multi-task learning, interpretability mechanisms, and uncertainty metrics demonstrated potential to facilitate effective clinical implementation.
Added Value of Intraoperative Data for Predicting Postoperative Complications: Development and Validation of a MySurgeryRisk Extension
检验使用术前和术中数据输入功能与仅使用术前数据相比，预测术后并发症的准确度、区分度和精确度提高的假设。预测术后并发症的模型通常忽略重要的术中生理变化。结合术中生理数据可提高模型性能。这项回顾性队列分析包括5年期间在单一机构进行的52，529例住院手术。经验证的MySurgeryRisk平台中的随机森林机器学习模型使用电子健康记录数据和患者社区特征对住院期间的三种术后并发症和死亡率进行了患者级别的预测。对于每个结果，一个模型仅使用术前数据进行训练，另一个模型使用术前和术中数据进行训练。模型进行了比较的准确性，歧视（表示为AUROC），精密度（表示为AUPRC），和重新分类指数（NRI）。结合术前和术中数据的机器学习模型在预测所有三种术后并发症方面比仅使用术前数据的模型具有更高的准确性、区分度和精确度（重症监护室住院时间> 48小时，机械通气> 48小时，神经系统并发症包括谵妄）和住院死亡率（准确度：88% vs. 77%，AUROC：0.93 vs. 0.87，AUPRC：0.21 vs. 0.15）。并发症的总体重新分类改善为2.9 - 10.0%，住院死亡率为11.2%。对术前和术中数据的分析显著提高了机器学习模型预测术后并发症的准确性、区分度和精确度。
Federated Learning with Multi-Partner OneFlorida+ Consortium Data for Predicting Major Postoperative Complications
Intelligent Clinical Care Center, University of Florida（佛罗里达大学智能临床护理中心）；Department of Medicine, Division of Nephrology, Hypertension, and Renal Transplantation, University of Florida（佛罗里达大学医学系肾病、高血压和肾移植分部）；Department of Surgery, University of Florida（佛罗里达大学外科系）；Department of Biomedical Engineering, University of Florida（佛罗里达大学生物医学工程系）
Self-explaining Hierarchical Model for Intraoperative Time Series
McKelvey School of Engineering, Washington University in St. Louis（圣路易斯华盛顿大学麦凯尔维工程学院）；School of Medicine, Washington University in St. Louis（圣路易斯华盛顿大学医学院）
Automated Prediction of Postoperative Pancreatic Fistula Using Preoperative Computed Tomography
Department of Surgery, Mayo Clinic, Rochester, MN, USA（梅奥诊所外科部，罗切斯特，明尼苏达州，美国）；Department of Surgery, University of Auckland, Auckland, NZ（奥克兰大学外科部，奥克兰，新西兰）；Kern Center for the Science of Health Care Delivery, Mayo Clinic, Rochester, MN, USA（健康照护科学中心，梅奥诊所，罗切斯特，明尼苏达州，美国）；Department of Artificial Intelligence（人工智能部）
Predicting Postoperative Nausea And Vomiting Using Machine Learning: A Model Development and Validation Study
Ariel University（阿里埃勒大学）；Holon Institute of Technology（霍隆理工学院）；Sheba Medical Center（示巴医疗中心）；Tel-Aviv University（特拉维夫大学）；University College London（伦敦大学学院）
Predicting Postoperative Stroke in Elderly SICU Patients: An Interpretable Machine Learning Model Using MIMIC Data
Department of Industrial and Systems Engineering, University of Southern California（1 工业与系统工程系，南加州大学）；Colorectal Research Center, Iran University of Medical Sciences（2 肠道癌研究中心，伊朗医学科学大学）；Department of Health Science, California State University, Long Beach (CSULB)（3 健康科学系，长滩加州州立大学（CSULB））；Department of Industrial and Manufacturing Engineering, California State Polytechnic University, Pomona（4 工业与制造工程系，庞纳加州州立理工大学）
A Self-Adaptive Frequency Domain Network for Continuous Intraoperative Hypotension Prediction
Jiangsu Key Laboratory of Intelligent Medical Image Computing, School of Artificial Intelligence, Nanjing University of Information Science（江苏智能医学图像计算重点实验室，人工智能学院，信息科学与技术大学）；Department of Anesthesiology, Surgery and Pain Management, School of Medicine, Zhongda Hospital, Southeast University, China（麻醉科、外科与疼痛管理部，医学院，东南大学中大医院，中国）
Transformer representation learning is necessary for dynamic multi-modal physiological data on small-cohort patients
Nanjing University（南京大学）；The Affiliated Drum Tower Hospital of Nanjing University Medical School（南京大学医学院附属鼓楼医院）；National Laboratory of Solid State Microstructure（固体微结构物理国家重点实验室）
本文提出结合 Transformer 表征与传统机器学习的 POD 预测框架，利用多模态生理数据构建首个两类患者数据集，证明多模态 Transformer 表征可提升术后 1–3 天谵妄诊断性能。
Explainable Machine Learning for ICU Readmission Prediction
School of Chemistry and Molecular Biosciences, The University of Queensland（昆士兰大学化学与分子生物科学学院）；Systems and Computational Biology, Bio21 Institute, The University of Melbourne（墨尔本大学生物21研究所系统与计算生物学系）；Computational Biology and Clinical Informatics, Baker Heart and Diabetes Institute（贝克心脏与糖尿病研究所计算生物学与临床信息学系）；Baker Department of Cardiometabolic Health, The University of Melbourne（墨尔本大学贝克代谢健康系）；Faculty of Medicine, Dentistry and Health Sciences, The University of Melbourne（墨尔本大学医学、牙医学与健康科学学院）；The Department of Surgery, St. Vincent’s Hospital Melbourne（墨尔本圣文森特医院外科）；Centre for the Digital Transformation of Health, School of Computing and Information Systems, The University of Melbourne（墨尔本大学计算与信息系统学院健康数字化转型中心）；The Department of Anaesthesia and Perioperative Medicine, Monash Health（莫纳什医疗中心麻醉与围手术医学科）；The Intensive Care Unit at Peninsula Private Hospital, Ramsay Healthcare（拉姆齐医疗半岛私立医院重症监护室）；The Department of Intensive Care Medicine, The Alfred Hospital（阿尔弗雷德医院重症医学科）；The Department of Anaesthesiology and Perioperative Medicine, The Alfred Hospital（阿尔弗雷德医院麻醉与围手术医学科）
