Automated Machine Learning for the Early Prediction of the Severity of Acute Pancreatitis in Hospitals

被引:19
作者
Yin, Minyue [1 ]
Zhang, Rufa [2 ]
Zhou, Zhirun [3 ]
Liu, Lu [1 ]
Gao, Jingwen [1 ]
Xu, Wei [1 ]
Yu, Chenyan [1 ]
Lin, Jiaxi [1 ]
Liu, Xiaolin [1 ]
Xu, Chunfang [1 ]
Zhu, Jinzhou [1 ]
机构
[1] Soochow Univ, Affiliated Hosp 1, Dept Gastroenterol, Suzhou, Peoples R China
[2] Soochow Univ, Changshu Hosp 1, Dept Gastroenterol, Suzhou, Peoples R China
[3] Soochow Univ, Affiliated Hosp 2, Dept Obstet & Gynecol, Suzhou, Peoples R China
关键词
automated machine learning; logistic regression analysis; severe acute pancreatitis; predictive models; artificial intelligence; PERSISTENT ORGAN FAILURE; CLINICAL SCORING SYSTEMS; PLEURAL EFFUSION; PROGNOSIS; OUTCOMES;
D O I
10.3389/fcimb.2022.886935
中图分类号
R392 [医学免疫学]; Q939.91 [免疫学];
学科分类号
100102 ;
摘要
BackgroundMachine learning (ML) algorithms are widely applied in building models of medicine due to their powerful studying and generalizing ability. This study aims to explore different ML models for early identification of severe acute pancreatitis (SAP) among patients hospitalized for acute pancreatitis. MethodsThis retrospective study enrolled patients with acute pancreatitis (AP) from multiple centers. Data from the First Affiliated Hospital and Changshu No. 1 Hospital of Soochow University were adopted for training and internal validation, and data from the Second Affiliated Hospital of Soochow University were adopted for external validation from January 2017 to December 2021. The diagnosis of AP and SAP was based on the 2012 revised Atlanta classification of acute pancreatitis. Models were built using traditional logistic regression (LR) and automated machine learning (AutoML) analysis with five types of algorithms. The performance of models was evaluated by the receiver operating characteristic (ROC) curve, the calibration curve, and the decision curve analysis (DCA) based on LR and feature importance, SHapley Additive exPlanation (SHAP) Plot, and Local Interpretable Model Agnostic Explanation (LIME) based on AutoML. ResultsA total of 1,012 patients were included in this study to develop the AutoML models in the training/validation dataset. An independent dataset of 212 patients was used to test the models. The model developed by the gradient boost machine (GBM) outperformed other models with an area under the ROC curve (AUC) of 0.937 in the validation set and an AUC of 0.945 in the test set. Furthermore, the GBM model achieved the highest sensitivity value of 0.583 among these AutoML models. The model developed by eXtreme Gradient Boosting (XGBoost) achieved the highest specificity value of 0.980 and the highest accuracy of 0.958 in the test set. ConclusionsThe AutoML model based on the GBM algorithm for early prediction of SAP showed evident clinical practicability.
引用
收藏
页数:11
相关论文
共 34 条
[1]   Establishing Machine Learning Models to Predict Curative Resection in Early Gastric Cancer with Undifferentiated Histology: Development and Usability Study [J].
Bang, Chang Seok ;
Ahn, Ji Yong ;
Kim, Jie-Hyun ;
Kim, Young-Il ;
Choi, Il Ju ;
Shin, Woon Geon .
JOURNAL OF MEDICAL INTERNET RESEARCH, 2021, 23 (04)
[2]   Classification of acute pancreatitis-2012: revision of the Atlanta classification and definitions by international consensus [J].
Banks, Peter A. ;
Bollen, Thomas L. ;
Dervenis, Christos ;
Gooszen, Hein G. ;
Johnson, Colin D. ;
Sarr, Michael G. ;
Tsiotos, Gregory G. ;
Vege, Santhi Swaroop .
GUT, 2013, 62 (01) :102-111
[3]   A practical guide to multiple imputation of missing data in nephrology [J].
Blazek, Katrina ;
van Zwieten, Anita ;
Saglimbene, Valeria ;
Teixeira-Pinto, Armando .
KIDNEY INTERNATIONAL, 2021, 99 (01) :68-74
[4]   A Comparative Evaluation of Radiologic and Clinical Scoring Systems in the Early Prediction of Severity in Acute Pancreatitis [J].
Bollen, Thomas L. ;
Singh, Vikesh K. ;
Maurer, Rie ;
Repas, Kathryn ;
van Es, Hendrik W. ;
Banks, Peter A. ;
Mortele, Koenraad J. .
AMERICAN JOURNAL OF GASTROENTEROLOGY, 2012, 107 (04) :612-619
[5]   The association of parameters of body composition and laboratory markers with the severity of hypertriglyceridemia-induced pancreatitis [J].
Chen, Lifang ;
Huang, Yingbao ;
Yu, Huajun ;
Pan, Kehua ;
Zhang, Zhao ;
Man, Yi ;
Hu, Dingyuan .
LIPIDS IN HEALTH AND DISEASE, 2021, 20 (01)
[6]   Machine Learning in Medicine [J].
Deo, Rahul C. .
CIRCULATION, 2015, 132 (20) :1920-1930
[7]   Management of Severe Acute Pancreatitis: An Update [J].
Gliem, Nina ;
Ammer-Herrmenau, Christoph ;
Ellenrieder, Volker ;
Neesse, Albrecht .
DIGESTION, 2021, 102 (04) :503-507
[8]   Artificial intelligence in sepsis early prediction and diagnosis using unstructured data in healthcare [J].
Goh, Kim Huat ;
Wang, Le ;
Yeow, Adrian Yong Kwang ;
Poh, Hermione ;
Li, Ke ;
Yeow, Joannas Jie Lin ;
Tan, Gamaliel Yu Heng .
NATURE COMMUNICATIONS, 2021, 12 (01)
[9]   Red cell distribution width and red cell distribution width to total serum calcium ratio as major predictors of severity and mortality in acute pancreatitis [J].
Gravito-Soares, Marta ;
Gravito-Soares, Elisa ;
Gomes, Dario ;
Almeida, Nuno ;
Tome, Luis .
BMC GASTROENTEROLOGY, 2018, 18
[10]   Assessment of the Prognostic Value of Certain Acute-Phase Proteins and Procalcitonin in the Prognosis of Acute Pancreatitis [J].
Gurda-Duda, Anna ;
Kusnierz-Cabala, Beata ;
Nowak, Wojciech ;
Naskalski, Jerzy W. ;
Kulig, Jan .
PANCREAS, 2008, 37 (04) :449-453