早期偵測敗血症不良結果：利用深度學習與模糊支持向量機

Early detection of sepsis utilizing deep learning and fuzzy support vector machine

龔千芬(Chien-Feng Kung)；龔嘉德(Chia-Te Kung)；蘇志民(Chih-Min Su)；郝沛毅(Pei-Yi Hao)；林奕儒(Yi-Ju Lin)

敗血症早期預測 ； 深度學習 ； 模糊支持向量機 ； 臨床決策支持系統 ； 醫學資訊學 ； Sepsis early prediction ； deep learning ； fuzzy support vector machines ； clinical decision support systems ； medical informatics

資訊管理學報

28卷4期（2021 / 10 / 31）

445 - 476

繁體中文

敗血症是全世界的主要死亡原因，敗血性休克的死亡率高達50％。根據世界衛生組織估計，每年有超過600萬人死於敗血症，早期診斷和治療可以預防大多數的敗血性休克發病與死亡，但是，目前缺乏可靠的早期敗血症智能預測系統。時至今日，在大數據分析的快速發展和重症監護室的豐富醫療數據不斷累積的推波助瀾下，讓人們對於開發智能模型來早期預測敗血性休克等急性醫療狀況產生了極大的興趣。本研究開發一套嶄新的智能敗血症早期預測技術，它包含長短期記憶體，卷積神經網路、完全連接網路與模糊支持向量機，以實現敗血症的早期預測。我們使用2010至2018年在長庚醫院接受醫療的17歲以上病患的電子健康記錄，提取的數據包含靜態特徵集，例如人口統計數據和過去病史；以及動態特徵集，例如帶有時間戳的生命徵象與生物標記。本研究提出一個混合的深度神經網路模型來自動學習關鍵特徵，第一個組件是卷積神經網路，它可以獲得動態信息的局部特徵。第二個組件是完全連接的神經網路，它可以擷取靜態信息內的隱含特徵，此外，長短期記憶體用於補抓動態信息的時間依賴性特徵，最後，深度模型學習得到的特徵將餵入到嶄新的模糊孿生支持向量機中，以預測敗血症不良結果。本研究提出的智能系統可以提前28天預測敗血症發作，這將為減輕敗血症發作的風險，提供早期解決方案。

Sepsis is a leading cause of death over the world and septic shock, the most severe complication of sepsis, reaches a mortality rate as high as 50%. The World Health Organization estimates that more than six million people die of sepsis annually, and early diagnosis and treatment can prevent most morbidity and mortality. However, reliable and intelligent systems for predicting sepsis are scarce. The rapid development in big data analytics and the data-rich environment of intensive care units has generated great interest in developing models to predict acute medical conditions such as septic shock. This study proposes a novel technique that combines long short-term memory (LSTM), convolutional neural network (CNN), fully connected neural network, and fuzzy twin support vector machine to achieve early prediction of sepsis. We used data from the electronic health records (EHRs) of all patients above 17 years old admitted to the medical ICU at Chang-Geng Memorial Hospital between January 2010 and December 2018. The extracted data contained sets of static features, such as demographic and clinical information, and temporal features such as time-stamped vital signs. In this study, we propose a general deep neural network framework that incorporates two additional components with the aim of improving LSTM to automatically extract important features. The first component, a CNN, is added before LSTM to obtain local characteristics of EHRs. The second component, a fully connected neural network, introduces static information (e.g., age) to LSTM. Finally, a LSTM is applied to handle dynamic information (e.g., the lab result). The features learned by the deep learning model are fed to a novel fuzzy twin support vector to predict sepsis onset in patients admitted to an intensive care unit. Using EHRs data, sepsis onset can be predicted up to 28 d in advance. Our findings will offer an early solution for mitigating the risk of sepsis onset.

1. Bhandary, A.,Prabhu, G.A.,Rajinikanth, V.,Thanaraj, K.P.,Satapathy, S.C.,Robbins, D.E.,Shasky, C.,Zhang, Y.-D.,Tavares, J.M.R.S.,Raja, N.S.M.(2020).Deep-Learning Framework to Detect Lung Abnormality - A study with Chest X-Ray and Lung CT Scan Images.Pattern Recognition Letters,129,271-278.
2. Demšar, J.(2006).Statistical Comparisons of Classifiers over Multiple Data Sets.Journal of Machine Learning Research,7(1),1-30.
3. Despins, L.A.(2017).Automated detection of sepsis using electronic medical record data: a systematic review.Journal for Healthcare Quality,39(6),322-333.
4. Doi, K.,Yuen, P.S.,Eisner, C.,Hu, X.,Leelahavanichkul, A.,Schnermann, J.,Star, R.A.(2009).Reduced production of creatinine limits its use as marker of kidney injury in sepsis.Journal of the American Society of Nephrology: JASN,20(6),1217-1221.
5. Fairchild, K.D.,O'Shea, T.M.(2010).Heart rate characteristics: physiomarkers for detection of late-onset neonatal sepsis.Clinics in Perinatology,37(3),581-598.
6. Froon, A.H.,Bemelmans, M.H.,Greve, J.W.,van der Linden, C.J.,Buurman, W.A.(1994).Increased plasma concentrations of soluble tumor necrosis factor receptors in sepsis syndrome: correlation with plasma creatinine values.Critical Care Medicine,22(5),803-809.
7. Gultepe, E.,Green, J.P.,Nguyen, H.,Adams, J.,Albertson, T.,Tagkopoulos, I.(2014).From vital signs to clinical outcomes for patients with sepsis: a machine learning basis for a clinical decision support system.Journal of the American Medical Informatics Association,21(2),315-325.
8. Hao, P.-Y.,Kung, C.-F.,Chang, C.-Y.,Ou, J.-B.(2020).Predicting Stock Price Trends Based on Financial News Articles and Using a Novel Twin Support Vector Machine with Fuzzy Hyperplane.Applied Soft Computing
9. Hatfield, K.M.,Dantes, R.B.,Baggs, J.,Sapiano, M.R.P.,Fiore, A.E.,Jernigan, J.A.,Epstein, L.(2018).Assessing variability in hospital-level mortality among US medicare beneficiaries with hospitalizations for severe sepsis and septic shock.Critical Care Medicine,46(11),1753-1760.
10. Hinton, G.E.,Salakhutdinov, R.R.(2006).Reducing the dimensionality of data with neural networks.Science,313,504-507.
11. Huang, F.J.,LeCun, Y.(2006).Large-scale Learning with SVM and Convolutional for Generic Object Categorization.2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition,New York, USA:
12. Islam, M.M.,Nasrin, T.,Walther, B.A.,Wu, C.C.,Yang, H.C.,Li, Y.C.(2019).Prediction of sepsis patients using machine learning approach: a meta-analysis.Computer Methods and Programs in Biomedicine,170,1-9.
13. Kok, C.,Jahmunah, V.,Oh, S.L.,Zhou, X.,Gururajan, R.,Tao, X.,Cheong, K.H.,Gururajan, R.,Molinari, F.,Acharya, U.R.(2020).Automated prediction of sepsis using temporal convolutional network.Computers in Biology and Medicine,127(103957),1-10.
14. Kumar, A.,Roberts, D.,Wood, K.E.,Light, B.,Parrillo, J.E.,Sharma, S.,Suppes, R.,Feinstein, D.,Zanotti, S.,Taiberg, L.,Gurka, D.,Kumar, A.,Cheang, M.(2006).Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock.Critical Care Medicine,34(6),1589-1596.
15. Lauritsen, S.M.,Kalør, M.E.,Kongsgaard, E.L.,Lauritsen, K.M.,Jørgensen, M.J.,Lange, J.,Thiesson, B.(2020).Early detection of sepsis utilizing deep learning on electronic health record event sequences.Artificial Intelligence in Medicine,104(1820),1-11.
16. Lin, C.,Zhang, Y.,Ivy, J.,Capan, M.,Arnold, R.,Huddleston, J.M.,Chi, M.(2018).Early Diagnosis and Prediction of Sepsis Shock by Combining Static and Dynamic Information Using Convolutional-LSTM.2018 IEEE International Conference on Healthcare Informatics (ICHI),New York:
17. Mao, Q.,Jay, M.,Hoffman, J.L.,Calvert, J.,Barton, C.,Shimabukuro, D.,Shieh, L.,Chettipally, U.,Fletcher, G.,Kerem, Y.,Zhou, Y.,Das, R.(2018).Multicentre validation of a sepsis prediction algorithm using only vital sign data in the emergency department, general ward and ICU.BMJ Open,8(1),e017833.
18. Nemati, S.,Holder, A.,Razmi, F.,Stanley, M.D.,Clifford, G.D.,Buchman, T.G.(2018).An interpretable machine learning model for accurate prediction of sepsis in the ICU.Critical Care Medicine,46(4),547-553.
19. Nguyen, H.B.,Rivers, E.P.,Knoblich, B.P.,Jacobsen, G.,Muzzin, A.,Ressler, J.A.,Tomlanovich, M.C.(2004).Early lactate clearance is associated with improved outcome in severe sepsis and septic shock.Critical Care Medicine,32(8),1637-1642.
20. Paoli, C.J.,Reynolds, M.A.,Sinha, M.,Gitlin, M.,Crouser, E.(2018).Epidemiology and costs of sepsis in the United States-an analysis based on timing of diagnosis and severity level.Critical Care Medicine,46(12),1889-1897.
21. Pierrakos, C.,Vincent, J.-L.(2010).Sepsis biomarkers: a review.Critical Care,14(R15),1-18.
22. Rafiei, A.,Rezaee, A.,Hajati, F.,Gheisari, S.,Golzan, M.(2021).SSP: Early prediction of sepsis using fully connected LSTM-CNN model.Computers in Biology and Medicine,128(104110),1-10.
23. Saqib, M.,Sha, Y.,Wang, M.D.(2018).Early Prediction of Sepsis in EMR Records Using Traditional ML Techniques and Deep Learning LSTM Networks.40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)
24. Shimabukuro, D.W.,Barton, C.W.,Feldman, M.D.,Mataraso, S.J.,Das, R.(2017).Effect of a machine learning-based severe sepsis prediction algorithm on patient survival and hospital length of stay: a randomised clinical trial.BMJ Open Respiratory Research,4(1),e000234.
25. Smith, G.B.,Prytherch, D.R.,Meredith, P.,Schmidt, P.E.,Featherstone, P. I.(2013).The ability of the National Early Warning Score (NEWS) to discriminate patients at risk of early cardiac arrest, unanticipated intensive care unit admission, and death.Resuscitation,84(4),465-470.
26. Taneja, I.,Reddy, B.,Damhorst, G.,Dave Zhao, S.,Hassan, U.,Price, Z.,Jensen, T.,Ghonge, T.,Patel, M.,Wachspress, S.,Winter, J.,Rappleye, M.,Smith, G.,Healey, R.,Ajmal, M.,Khan, M.,Patel, J.,Rawal, H.,Sarwar, R.,Zhu, R.(2017).Combining Biomarkers with EMR Data to Identify Patients in Different Phases of Sepsis.Scientific Reports,7(10800),1-12.
27. Tao, X.,Li, Q.,Ren, C.,Guo, W.,He, Q.,Liu, R.,Zou, J.(2020).Affinity and class probability-based fuzzy support vector machine for imbalanced data sets.Neural Networks,122,289-307.
28. Taylor, R.A.,Pare, J.,Venkatesh, A.,Mowafi, H.,Melnick, E.,Fleischman, W.,Hall, M.K.(2016).Prediction of in-hospital mortality in emergency department patients with sepsis: a local big data-driven, machine learning approach.Academic Emergency Medicine,23(3),269-278.
29. Torio, C.M.,Moore, B.J.(2016).National Inpatient Hospital Costs: the Most Expensive Conditions by Payer, 2013.Healthcare Cost and Utilization Project (HCUP) Statistical Briefs,204,2006-2016.
30. Xu, B.,Shirani, A.,Lo, D.,Alipour, M.A.(2018).Prediction of relatedness in stack overflow: deep learning vs. SVM: a reproducibility study.Proceedings of the 12th ACM/IEEE International Symposium on Empirical Software Engineering and Measurement (ESEM '18),New York, USA:
31. Ye, Y.,Tian, M.,Liu, Q.,Tai, H.-M.(2020).Pulmonary Nodule Detection Using V-Net and High-Level Descriptor Based SVM Classifier.IEEE Access,8,176033-176041.

1. 龔千芬,郝沛毅(2022)。融合深度神經網路與深層模糊孿生支持向量機於股價預測。資訊管理學報，29(4)，303-333。