Seung Hee Lim, Min Jeong Kim, Won Hyuk Choi, Jin Cheol Cheong, Jong Wan Kim, Kyung Joo Lee, Jun Ho Park
{"title":"可解释的机器学习使用围手术期系列实验室结果预测腹膜炎诱导败血症患者的术后死亡率。","authors":"Seung Hee Lim, Min Jeong Kim, Won Hyuk Choi, Jin Cheol Cheong, Jong Wan Kim, Kyung Joo Lee, Jun Ho Park","doi":"10.4174/astr.2023.105.4.237","DOIUrl":null,"url":null,"abstract":"<p><strong>Purpose: </strong>Sepsis is one of the most common causes of death after surgery. Several conventional scoring systems have been developed to predict the outcome of sepsis; however, their predictive power is insufficient. The present study applies explainable machine-learning algorithms to improve the accuracy of predicting postoperative mortality in patients with sepsis caused by peritonitis.</p><p><strong>Methods: </strong>We performed a retrospective analysis of data from demographic, clinical, and laboratory analyses, including the delta neutrophil index (DNI), WBC and neutrophil counts, and CRP level. Laboratory data were measured before surgery, 12-36 hours after surgery, and 60-84 hours after surgery. The primary study output was the probability of mortality. The areas under the receiver operating characteristic curves (AUCs) of several machine-learning algorithms using the Sequential Organ Failure Assessment (SOFA) and Simplified Acute Physiology Score (SAPS) 3 models were compared. 'SHapley Additive exPlanations' values were used to indicate the direction of the relationship between a variable and mortality.</p><p><strong>Results: </strong>The CatBoost model yielded the highest AUC (0.933) for mortality compared to SAPS3 and SOFA (0.860 and 0.867, respectively). Increased DNI on day 3, septic shock, use of norepinephrine therapy, and increased international normalized ratio on day 3 had the greatest impact on the model's prediction of mortality.</p><p><strong>Conclusion: </strong>Machine-learning algorithms increase the accuracy of predicting postoperative mortality in patients with sepsis caused by peritonitis.</p>","PeriodicalId":8071,"journal":{"name":"Annals of Surgical Treatment and Research","volume":"105 4","pages":"237-244"},"PeriodicalIF":1.2000,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10613826/pdf/","citationCount":"0","resultStr":"{\"title\":\"Explainable machine learning using perioperative serial laboratory results to predict postoperative mortality in patients with peritonitis-induced sepsis.\",\"authors\":\"Seung Hee Lim, Min Jeong Kim, Won Hyuk Choi, Jin Cheol Cheong, Jong Wan Kim, Kyung Joo Lee, Jun Ho Park\",\"doi\":\"10.4174/astr.2023.105.4.237\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><strong>Purpose: </strong>Sepsis is one of the most common causes of death after surgery. Several conventional scoring systems have been developed to predict the outcome of sepsis; however, their predictive power is insufficient. The present study applies explainable machine-learning algorithms to improve the accuracy of predicting postoperative mortality in patients with sepsis caused by peritonitis.</p><p><strong>Methods: </strong>We performed a retrospective analysis of data from demographic, clinical, and laboratory analyses, including the delta neutrophil index (DNI), WBC and neutrophil counts, and CRP level. Laboratory data were measured before surgery, 12-36 hours after surgery, and 60-84 hours after surgery. The primary study output was the probability of mortality. The areas under the receiver operating characteristic curves (AUCs) of several machine-learning algorithms using the Sequential Organ Failure Assessment (SOFA) and Simplified Acute Physiology Score (SAPS) 3 models were compared. 'SHapley Additive exPlanations' values were used to indicate the direction of the relationship between a variable and mortality.</p><p><strong>Results: </strong>The CatBoost model yielded the highest AUC (0.933) for mortality compared to SAPS3 and SOFA (0.860 and 0.867, respectively). Increased DNI on day 3, septic shock, use of norepinephrine therapy, and increased international normalized ratio on day 3 had the greatest impact on the model's prediction of mortality.</p><p><strong>Conclusion: </strong>Machine-learning algorithms increase the accuracy of predicting postoperative mortality in patients with sepsis caused by peritonitis.</p>\",\"PeriodicalId\":8071,\"journal\":{\"name\":\"Annals of Surgical Treatment and Research\",\"volume\":\"105 4\",\"pages\":\"237-244\"},\"PeriodicalIF\":1.2000,\"publicationDate\":\"2023-10-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10613826/pdf/\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Annals of Surgical Treatment and Research\",\"FirstCategoryId\":\"3\",\"ListUrlMain\":\"https://doi.org/10.4174/astr.2023.105.4.237\",\"RegionNum\":4,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"2023/9/27 0:00:00\",\"PubModel\":\"Epub\",\"JCR\":\"Q3\",\"JCRName\":\"SURGERY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Annals of Surgical Treatment and Research","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.4174/astr.2023.105.4.237","RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2023/9/27 0:00:00","PubModel":"Epub","JCR":"Q3","JCRName":"SURGERY","Score":null,"Total":0}
Explainable machine learning using perioperative serial laboratory results to predict postoperative mortality in patients with peritonitis-induced sepsis.
Purpose: Sepsis is one of the most common causes of death after surgery. Several conventional scoring systems have been developed to predict the outcome of sepsis; however, their predictive power is insufficient. The present study applies explainable machine-learning algorithms to improve the accuracy of predicting postoperative mortality in patients with sepsis caused by peritonitis.
Methods: We performed a retrospective analysis of data from demographic, clinical, and laboratory analyses, including the delta neutrophil index (DNI), WBC and neutrophil counts, and CRP level. Laboratory data were measured before surgery, 12-36 hours after surgery, and 60-84 hours after surgery. The primary study output was the probability of mortality. The areas under the receiver operating characteristic curves (AUCs) of several machine-learning algorithms using the Sequential Organ Failure Assessment (SOFA) and Simplified Acute Physiology Score (SAPS) 3 models were compared. 'SHapley Additive exPlanations' values were used to indicate the direction of the relationship between a variable and mortality.
Results: The CatBoost model yielded the highest AUC (0.933) for mortality compared to SAPS3 and SOFA (0.860 and 0.867, respectively). Increased DNI on day 3, septic shock, use of norepinephrine therapy, and increased international normalized ratio on day 3 had the greatest impact on the model's prediction of mortality.
Conclusion: Machine-learning algorithms increase the accuracy of predicting postoperative mortality in patients with sepsis caused by peritonitis.
期刊介绍:
Manuscripts to the Annals of Surgical Treatment and Research (Ann Surg Treat Res) should be written in English according to the instructions for authors. If the details are not described below, the style should follow the Uniform Requirements for Manuscripts Submitted to Biomedical Journals: Writing and Editing for Biomedical Publications available at International Committee of Medical Journal Editors (ICMJE) website (http://www.icmje.org).