An AI-driven machine learning approach identifies risk factors associated with 30-day mortality following total aortic arch replacement combined with stent elephant implantation

An AI-driven machine learning approach identifies risk factors associated with 30-day mortality following total aortic arch replacement combined with stent elephant implantation | medRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-P4HH5NV'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search An AI-driven machine learning approach identifies risk factors associated with 30-day mortality following total aortic arch replacement combined with stent elephant implantation Shuai Zhang , Lulu Li , Jingyu Wang , Yuan Li , Cuntao Yu , Xiaogang Sun , Jing Sun , Xiangyang Qian doi: https://doi.org/10.1101/2025.03.06.25323548 Shuai Zhang 1 Department of Cardiovascular Surgery at Fuwai Hospital, affiliated with the Chinese Academy of Medical Sciences & Peking Union Medical College/National Center for Cardiovascular Diseases, China ; 167 North Lishi Road, Xicheng District, Beijing 100037 in the Xicheng District of Beijing, China Find this author on Google Scholar Find this author on PubMed Search for this author on this site Lulu Li 1 Department of Cardiovascular Surgery at Fuwai Hospital, affiliated with the Chinese Academy of Medical Sciences & Peking Union Medical College/National Center for Cardiovascular Diseases, China ; 167 North Lishi Road, Xicheng District, Beijing 100037 in the Xicheng District of Beijing, China Find this author on Google Scholar Find this author on PubMed Search for this author on this site Jingyu Wang 2 Key Laboratory of Cardiovascular Epidemiology, located in the Department of Epidemiology at Fuwai Hospital, is affiliated with the State Key Laboratory of Cardiovascular Disease, Chinese Academy of Medical Sciences & Peking Union Medical College/National Center for Cardiovascular Diseases . China , Beijing, Xicheng District, 167 North Lishi Road, 100037 Find this author on Google Scholar Find this author on PubMed Search for this author on this site Yuan Li 1 Department of Cardiovascular Surgery at Fuwai Hospital, affiliated with the Chinese Academy of Medical Sciences & Peking Union Medical College/National Center for Cardiovascular Diseases, China ; 167 North Lishi Road, Xicheng District, Beijing 100037 in the Xicheng District of Beijing, China Find this author on Google Scholar Find this author on PubMed Search for this author on this site Cuntao Yu 1 Department of Cardiovascular Surgery at Fuwai Hospital, affiliated with the Chinese Academy of Medical Sciences & Peking Union Medical College/National Center for Cardiovascular Diseases, China ; 167 North Lishi Road, Xicheng District, Beijing 100037 in the Xicheng District of Beijing, China Find this author on Google Scholar Find this author on PubMed Search for this author on this site Xiaogang Sun 1 Department of Cardiovascular Surgery at Fuwai Hospital, affiliated with the Chinese Academy of Medical Sciences & Peking Union Medical College/National Center for Cardiovascular Diseases, China ; 167 North Lishi Road, Xicheng District, Beijing 100037 in the Xicheng District of Beijing, China Find this author on Google Scholar Find this author on PubMed Search for this author on this site Jing Sun 1 Department of Cardiovascular Surgery at Fuwai Hospital, affiliated with the Chinese Academy of Medical Sciences & Peking Union Medical College/National Center for Cardiovascular Diseases, China ; 167 North Lishi Road, Xicheng District, Beijing 100037 in the Xicheng District of Beijing, China Find this author on Google Scholar Find this author on PubMed Search for this author on this site Xiangyang Qian 1 Department of Cardiovascular Surgery at Fuwai Hospital, affiliated with the Chinese Academy of Medical Sciences & Peking Union Medical College/National Center for Cardiovascular Diseases, China ; 167 North Lishi Road, Xicheng District, Beijing 100037 in the Xicheng District of Beijing, China Find this author on Google Scholar Find this author on PubMed Search for this author on this site For correspondence: m13701097213{at}163.com Abstract Full Text Info/History Metrics Data/Code Preview PDF Abstract Objectives During emergency surgery, patients with acute type A aortic dissection (ATAAD) experience unfavorable outcomes throughout their hospital stay. The combination of total aortic arch replacement (TAR) and frozen elephant trunk (FET) implantation has become a dependable choice for surgical treatment. The objective of this research was to utilize a machine learning technique based on artificial intelligence to detect the factors that increase the risk of mortality within 30 days after surgery in patients who undergo TAR in combination with FET. Methods From January 2015 to December 2020, a total of 640 patients with ATAAD who underwent TAR and FET were included in this study. The subjects were divided into a test group and a validation group in a random manner, with a ratio of 7 to 3. The objective of our research was to create predictive models by employing different supervised machine learning techniques, such as XGBoost, logistic regression, support vector machine (SVM), and random forest (RF), to assess and compare their respective performances. Furthermore, we employed SHapley Additive exPlanation (SHAP) measures to allocate interpretive attributional values. Results Among all the patients, 37 (5.78%) experienced perioperative mortality. Subsequently, a total 50 of 10 highly associated variables were selected for model construction. By implementing the new method, the AUC value significantly improved from 0.6981 using the XGBoost model to 0.8687 with the PSO-ELM-FLXGBoost model. Conclusion In this study, machine learning methods were successfully established to predict ATAAD perioperative mortality, enabling the optimization of postoperative treatment strategies to minimize the postoperative complications following cardiac surgeries. 1. Introduction Acute Type A Aortic Dissection (ATAAD) is a severe condition linked to significant fatality rates (16−19% for surgical mortality) 1 – 2 . The mortality of ATAAD increases over time, with a rate of 1% per hour 3 . ATAAD is primarily treated with Total Aortic Arch Replacement (TAR) using deep hypothermic circulatory arrest. The approaches to surgically treating ATAAD lesions differ because of the complex and diverse characteristics of these lesions 4 . In contrast to Western nations, the most common surgery for ATAAD in China is TAR with the frozen elephant trunk (FET) method 5 . Nevertheless, despite the progress made in surgical methods and perioperative care, the fatality rate after surgical intervention remains elevated. Given its intricate nature and elevated risk, the success of TAR+FET largely depends on the patient’s condition before surgery and the specifics of the surgical process. Premature death is linked to numerous high-risk elements such as advanced age, inadequate pre-surgery organ perfusion, heart tamponade, low blood pressure, stroke after surgery, continuous renal replacement therapy, low cardiac output syndrome, and multiple organ dysfunction 6 . Hence, it is crucial to create suitable forecasting models for evaluating linked hazards and recognizing populations at high risk in clinical settings. Logistic regression is widely applied in the construction of various medical predictive models, serving as a linear model for solving binary classification problems. However, Logistic regression has several limitations, such as assuming a linear relationship between independent and dependent variables, sensitivity to multicollinearity, and challenges with imbalanced categories. Consequently, this study applies an improved machine learning method called RF-PSO-FLXGBoost to establish an accurate prediction model. 2. Methods 2.1. Study population and Definition From January 2015 to December 2020, a total of 1,113 individuals diagnosed with ATAAD and treated surgically at Fuwai Hospital were included in the study. A total of 123 patients who underwent isolated ascending aorta replacement and 147 patients who had ascending aorta with hemi-arch replacement were omitted from the final analysis. 203 chronic Type A aortic dissection patients were also excluded, remaining 640 individuals with TAR+FET. We gathered demographic details, pre-surgery risk elements, and crucial intraoperative data from all patients for examination. Approval was obtained (no.2020-1402) from the Ethics Committee of Fuwai hospital. The research involved minimal risk to patients, so patient informed consent was not required. The waiver has no negative impact on the participants’ rights and well-being. 2.2. Machine learning In order to compare the effect of different variables on the outcome and identify the key factors, we applied StandardScaler to standardize feature data, transforming the values of each feature into a standard normal distribution with a mean of 0 and a standard deviation of 1 (also known as the Z-score distribution). The variables that met the inclusion criteria were successively entered into each machine learning model, and the AUC test was performed on different models. The predictors were obtained according to the final improved algorithm, and the SHAP was used for interpretation. Figure 1 illustrates the entire procedure. Download figure Open in new tab Figure 1. Study Flow chart. Abbreviations: ELM, Extreme Learning Machine; PSO, Particle Swarm Optimization; FL, Focal Loss. 2.3. Data Pre-processing ELM 7 , also known as Extreme Learning Machine, is a type of feed-forward neural network that requires only the specification of the hidden layer’s neuron count. By utilizing random generation of input weights and hidden layer thresholds, acquiring optimal output weights through least squares, and fitting output values through stochastic mapping, this approach offers the benefits of exceptional efficiency and robust generalization capabilities. 2.4. Variable selection Our study chose appropriate variables for model building by multistage screening method. In the first stage, we eliminated variables that significantly unrelated to the outcome. Using Pearson’s correlation analysis, we also examined the multicollinearity between all variables, and selected applicable variables according to the previous research results. In the second stage, we applied XGBoost to generate feature importance of each variable retained in the first stage, indicating their predictive values for the model 8 – 9 . Ultimately, the top 10 variables with highest feature importance were selected as the predictive factors for the model. 3. Model derivation 3.1. Prediction model This study utilized the XGBoost technique and employed the latest ELM to replace the missing data. Additionally, PSO (Particle Swarm Optimization)and FL(Focal Loss) were employed for parameter adjustment, and the model underwent continuous optimization to achieve the best possible machine learning model 10 – 11 . 3.2. Evaluation of prediction model The validation group was employed to determine the accuracy of different machine learning models by computing the areas under the ROC curve. Moreover, in order to better interpret the association between variables and outcome, our study applied SHapley Additive exPlanation (SHAP) method to assign consistent attribution values to each variable in the model 12 .A variable’s contribution to risk prediction increases as the SHAP absolute value of the variable increases. 3.3. Statistical Analysis All charts were created using the R programming tool ( http://www.R-project.org ; Version 4.2.1). Statistical significance was determined by a p-value less than 0.05 or an adjusted p-value below 0.05. To compare the differences between the two groups, we conducted a parametric analysis using Student’s t-test and a non-parametric analysis with the Mann-Whitney U test. The p value was adjusted using the Benjamini-Hochberg method, and categorical variables were examined with the Chi-square test. Continuous variables that were normally distributed or nearly normal were represented using the mean and standard deviation. The median [IQR] was used to describe the continuous variables of skewness distribution. 4. Results 4.1. Training and validation of models From January 2015 to December 2020, 1113 consecutive patients treated surgically were admitted to the Fuwai Hospital. We selected patients according to the relevant criteria. After normalizing the data with z-scores, the patients were randomly split, assigning 70% to the training group and the other 30% to the testing group. For all variables, the differences between the training set and the test set were nonsignificant. Figure 2 displays the heatmap of the ten most significant variables linked to intensity in the initial stage (using XGBoost exclusively). Download figure Open in new tab Download figure Open in new tab Figure 2. (A) Hot map of the selected variables.(B) The hot map of the top ten variables related to intensity in first stage. ALT, Alanine aminotransferase; AST, Aspartate aminotransferase; BMI, body mass index; CABG, coronary artery bypass grafting; CK-MB, Creatine kinase-MB; COPD, Chronic obstructive pulmonary disease; CPB, cardiopulmonary bypass; EGFR, Estimated Glomerular Filtration Rate; LDH, lactate dehydrogenase; LVD, left ventricular internal diameter; LVEF, left ventricular ejection fraction; LYMPH, lymphocyte; NEUT, neutrophil; NLR, neutrophil to lymphocyte ratio; NYHA, New York Heart Association Classification; PLT, platelets; RBC, red blood cell; TEVAR, thoracic endovascular aortic repair; WBC, white blood cell. Of all the patients (mean age 46.78±10.12 years, females accounted for 21.09%) enrolled in this study, 83.91% had hypertension, 9.22% had Marfan syndrome. A total of 37 patients (5.78%) developed postoperative 30-Day mortality and were included in the outcome event collection. Table 1 provides a comprehensive overview of the demographic and perioperative data. View this table: View inline View popup Table 1 Patient characteristics and perioperative variables are presented in Table 1. 4.2. Model Evaluation and Interpretation Figure 3 displays the AUCs obtained by employing the enhanced PSO-ELM-FLXGBoost machine learning technique, utilizing all variables as input variables. The AUC (0.869) for the improved model was largest. Download figure Open in new tab Figure 3. By implementing new methods, the AUC value significantly improved from 0.6981 using the XGBoost model to 0.8687 with the PSO-ELM-FLXGBoost model. (A) The AUC (0.7514) for the ELM-FLXGBoost model was larger than that for the MF-XGBoost (0.7274) and XGBoost (0.6981). (B)The AUC (0.8687) for the PSO-ELM-FLXGBoost model was largest. Figure 4 displays the significance matrix chart for the PSO-FLXGBoost technique, indicating that the model’s top 10 influential factors include Gender, age, BMI, Lower limb ischemia, EGFR < 50 ml/min/1.73m 2 , ALT, LDH, D-Dimer, transfusion of red blood cells and Cardiopulmonary bypass time. Download figure Open in new tab Figure 4. Plotting the matrix of importance for the PSO-FLXGBoost technique. The model is influenced by the top 10 variables that are of utmost significance. The SHAP summary plot was utilized to determine the characteristics that had the greatest impact on the prediction model. We combine the feature value of the y-axis and the SHAP value of the x-axis, we can see that older age, longer CPB time, higher ALT, D-Dimer and EGFR can significantly increase the probability of perioperative mortality. SHAP analysis of all significant variables is shown in Figure 5 . Download figure Open in new tab Figure 5. SHAP analysis of top 10 significant variables. 5. Discussion According to data from the International Registry of Aortic Dissection (IRAD) cohort, the mean age of patients is 61.5 ± 14.6 years 13 . However, in China, the mean age of patients with ATAAD is significantly younger at 46.8 ± 12.1 years, which may be due to differences in aortic pathology and poor blood pressure control, resulting in a higher proportion of Chinese TAAA patients with Marfan syndrome and aortic dissection. Furthermore, a retrospective study of ATAAD patients revealed that TAR+FET is frequently employed in patients under 50 years of age 14 . This surgical approach is preferred by surgeons due to the longer life expectancy of Chinese patients. Despite the relative standardization of the TAR procedure and advancements in perioperative care, postoperative mortality following surgical intervention remains high. Accurate assessment of the condition’s severity and timely intervention can significantly enhance patient outcomes and reduce mortality rates. Thus, the development of effective predictive modeling is essential to evaluate associated risks and identify higher-risk populations within clinical settings. In our past work, we used a traditional logistic way to develop and validate a nomogram model and applied it in clinical situations 15 . The nomogram is a simple and effective tool in our clinical practice. However, no method is perfect. Regression model explain well but are not flexible enough, and machine learning methods are flexible but have limited explanatory power. Clinical predictive modeling as an intelligent tool for risk and benefit assessment, how we optimize the model is an important issue. With the development of machine learning predictive models, we note that ML leverages artificial intelligence to autonomously extract valuable information, identify potential patterns within large datasets, and develop robust risk models 16 . Compared to traditional analytical methods, ML demonstrates superior predictive strength and stability while addressing nonlinear and complex interactions between variables and outcomes. As a result, ML is increasingly employed across various medical domains, including diagnosis, prognostic prediction, treatment, and medical image analysis 17 . However, there is a notable lack of research specifically focused on the use of machine learning to evaluate the risk of perioperative mortality in individuals diagnosed with Acute Type A Aortic Dissection following TAR+FET. It is widely recognized that machine learning techniques are often used on extensive datasets. While the sample size in this research is adequate for TAR+FET, it is somewhat limited for machine learning algorithms, potentially leading to subpar model performance. To optimize the prediction, Particle Swarm Optimization (PSO), an intelligent algorithm that simulates bird behavior, was employed. Additionally, considering the small proportion of individuals with adverse outcomes, we employed Imbalance-XGBoost to improve the predictions. This involved applying PSO and the FL function to further enhance the algorithm. To better interpret the association between variables and outcomes, our study employed the SHAP (SHapley Additive exPlanations) method, which assigns consistent attribution values to each variable in the model 18 . Variables with higher SHAP absolute values made greater contributions to risk prediction. The PSO-ELM-FLXGBoost model outperformed the other models in terms of prediction, achieving the highest AUC among all the models tested. Moreover, this model demonstrated improved predictive power even with small datasets or imbalanced outcome events compared to the total sample size. We hope that this optimized PSO-ELM-FLXGBoost model can inform the construction of machine learning in clinical prediction models. In our study, machine learning methods were successfully established to predict ATAAD perioperative mortality in our cohort. But most importantly, The ultimate goal of medical predictive model construction is for clinical application. We hope to provide some ideas for future predictive modeling-assisted clinical decision making and believe that traditional regression modeling and machine learning approaches each have their own characteristics and can be better applied to clinical practice if they complement each other. We are looking forward to optimizing the model in the follow-up study to combine the traditional model with the machine learning approach for better clinical applications. 6. Limitations There are various constraints in this study that require attention. It is crucial to mention that this study is retrospective and conducted in a single center, which could lead to bias because of patients at a single center may not be broadly representative of the general population. Additionally, the effectiveness of machine learning algorithms may differ based on the dataset’s magnitude and the distribution of patient attributes. Moreover, it is crucial to acknowledge that this model lacks external validation, potentially restricting its applicability and reducing its accuracy in representing real-life situations. Hence, additional research is required to assess the feasibility of this model. 7. Conclusions To sum up, this research employed preoperative lab tests, clinical imaging records, and initial postoperative therapy results of individuals diagnosed with aortic dissection. Our team has effectively created and verified machine learning algorithms for forecasting the occurrence of death within 30 days after surgery in individuals diagnosed with acute type A aortic dissection. The death risk model suggested in this research is simple and easy to use, allowing medical practitioners to evaluate the risks associated with the condition and treatment, anticipate the probability of mortality in individuals with aortic dissection, direct the creation and execution of suitable and efficient treatment strategies, improve prognosis and survival rates, and ultimately decrease the rates of mortality. Data Availability Data cannot be shared publicly because of data sharing agreements and research ethics board protocols with participating hospitals.. Data are available from the Fuwai hospital Ethics Committee (contact via https://www.fuwai.com/News/Articles/Index/192 ) for researchers who meet the criteria for access to confidential data. Fundings This work was supported Beijing Municipal Science & Technology Commission (No. 2020-BKJ01); National High Level Hospital Clinical Research Funding (No.2023-GSP-GG-25). Conflict of interest The writers assert that they do not possess any recognized conflicting monetary concerns or personal associations that may have seemed to impact the research detailed in this document. Reference 1. ↵ Conzelmann LO , Weigang E , Mehlhorn U , Abugameh A , Hoffmann I , Blettner M , et al. Mortality in patients with acute aortic dissection type A: analysis of pre- and intraoperative risk factors from the German Registry for Acute Aortic Dissection Type A (GERAADA) . Eur J Cardiothorac Surg . 2016 ; 49 ( 2 ): e44 – 52 . doi: 10.1093/ejcts/ezv356 OpenUrl CrossRef PubMed 2. ↵ Inoue Y , Matsuda H , Uchida K , Komiya T , Koyama T , Yoshino H , et al. Analysis of Acute Type A Aortic Dissection in Japan Registry of Aortic Dissection (JRAD) . Ann Thorac Surg . 2020 ; 110 ( 3 ): 790 – 8 . doi: 10.1016/j.athoracsur.2019.12.051 OpenUrl CrossRef PubMed 3. ↵ Jassar AS , Sundt TM , 3rd. How should we manage type A aortic dissection? Gen Thorac Cardiovasc Surg . 2019 ; 67 ( 1 ): 137 – 45 . doi: 10.1007/s11748-018-0957-3 OpenUrl CrossRef PubMed 4. ↵ Qiu J , Luo X , Wu J , PanW , Chang Q , Qian X , et al. New aortic arch dissection classification: the fuwai classification . Front Cardiov Med . ( 2021 ) 8 : 710281 . doi: 10.3389/fcvm.2021.710281 OpenUrl CrossRef 5. ↵ Sun L , Qi R , Zhu J , Liu Y , Zheng J . Total arch replacement combined with stented elephant trunk implantation: a new “standard” therapy for type a dissection involving repair of the aortic arch? Circulation . ( 2011 ) 123 : 971 – 8 . doi: 10.1161/CIRCULATIONAHA.110.015081 OpenUrl Abstract / FREE Full Text 6. ↵ Carrel T , Sundt TM , 3rd, von Kodolitsch Y , Czerny M. Acute aortic dissection. Lancet . 2023 ; 401 ( 10378 ): 773 – 88 . doi: 10.1016/S0140-6736(22)01970-5 OpenUrl CrossRef PubMed 7. ↵ Tang J , Deng C , Huang GB . Extreme Learning Machine for Multilayer Perceptron . IEEE Trans Neural Netw Learn Syst . 2016 ; 27 ( 4 ): 809 – 21 . doi: 10.1109/TNNLS.2015.2424995 OpenUrl CrossRef PubMed 8. ↵ Mortazavi BJ , Bucholz EM , Desai NR , Huang C , Curtis JP , Masoudi FA , et al. Comparison of Machine Learning Methods With National Cardiovascular Data Registry Models for Prediction of Risk of Bleeding After Percutaneous Coronary Intervention . JAMA Netw Open . 2019 ; 2 ( 7 ): e196835 . doi: 10.1001/jamanetworkopen.2019.6835 OpenUrl CrossRef 9. ↵ Kilic A , Goyal A , Miller JK , Gjekmarkaj E , Tam WL , Gleason TG , et al. Predictive Utility of a Machine Learning Algorithm in Estimating Mortality Risk in Cardiac Surgery . Ann Thorac Surg . 2020 ; 109 ( 6 ): 1811 – 9 . doi: 10.1016/j.athoracsur.2019.09.049 OpenUrl CrossRef PubMed 10. ↵ Rusdah DA , Murfi H . XGBoost in handling missing values for life insurance risk prediction . SN Applied Sciences . 2020 ; 2 ( 8 ): 1 – 10 . DOI: 10.1007/s42452-020-3128-y OpenUrl CrossRef 11. ↵ Zhao T , Chen H , Bai Y , Zhao Y , Zhao S . A Hierarchical Ensemble Deep Learning Activity Recognition Approach with Wearable Sensors Based on Focal Loss . Int J Environ Res Public Health . 2022 ; 19 ( 18 ). doi: 10.3390/ijerph191811706 OpenUrl CrossRef 12. ↵ Silveira EC , Pretti SM , Santos BA , Correa CFS , Silva LM , Melo FFD . Prediction of hospital mortality in intensive care unit patients from clinical and laboratory data: A machine learning approach . World journal of critical care medicine . 2022 ( 005 ): 011 . doi: 10.5492/wjccm.v11.i5.317 OpenUrl CrossRef 13. ↵ Evangelista A , Isselbacher E , Bossone E , Gleason T , Eusanio M , Sechtem U , et al. Insights from the international registry of acute aortic dissection: a 20-year experience of collaborative clinical research . Circulation . ( 2018 ) 137 : 1846 – 60 . doi: 10.1161/CIRCULATIONAHA.117 . 031264 OpenUrl Abstract / FREE Full Text 14. ↵ Tamura K , Chikazawa G , Hiraoka A , Totsugawa T , Yoshitaka H . Characteristics and surgical results of acute type a aortic dissection in patients younger than 50 years of age . Ann Vasc Dis . ( 2019 ) 12 : 507 – 13 . doi: 10.3400/avd.oa.19-00033 OpenUrl CrossRef PubMed 15. ↵ Lin H , Chang Y , Guo H , Qian X , Sun X , Yu C . Prediction Nomogram for Postoperative 30-Day Mortality in Acute Type A Aortic Dissection Patients Receiving Total Aortic Arch Replacement With Frozen Elephant Trunk Technique . Front Cardiovasc Med . 2022 Jun 10 ; 9 : 905908 . doi: 10.3389/fcvm.2022.905908 . OpenUrl CrossRef 16. ↵ Al’Aref SJ , Anchouche K , Singh G , Slomka PJ , Kolli KK , Kumar A , et al. Clinical applications of machine learning in cardiovascular disease and its relevance to cardiac imaging . Eur Heart J . 2019 ; 40 ( 24 ): 1975 – 86 . doi: 10.1093/eurheartj/ehy404 OpenUrl CrossRef PubMed 17. ↵ Shameer K , Johnson KW , Glicksberg BS , Dudley JT , Sengupta PP . Machine learning in cardiovascular medicine: are we there yet? Heart . 2018 ; 104 ( 14 ): 1156 – 64 . doi: 10.1136/heartjnl-2017-311198 OpenUrl Abstract / FREE Full Text 18. ↵ Geisbusch S , Stefanovic A , Schray D , Oyfe I , Lin HM , Di Luozzo G , et al. A prospective study of growth and rupture risk of small-to-moderate size ascending aortic aneurysms . J Thorac Cardiovasc Surg . 2014 ; 147 ( 1 ): 68 – 74 . http://doi.org/10.1016/j.jtcvs.2013.06.030 OpenUrl PubMed View the discussion thread. Back to top Previous Next Posted March 10, 2025. Download PDF Data/Code Email Thank you for your interest in spreading the word about medRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following An AI-driven machine learning approach identifies risk factors associated with 30-day mortality following total aortic arch replacement combined with stent elephant implantation Message Subject (Your Name) has forwarded a page to you from medRxiv Message Body (Your Name) thought you would like to see this page from the medRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share An AI-driven machine learning approach identifies risk factors associated with 30-day mortality following total aortic arch replacement combined with stent elephant implantation Shuai Zhang , Lulu Li , Jingyu Wang , Yuan Li , Cuntao Yu , Xiaogang Sun , Jing Sun , Xiangyang Qian medRxiv 2025.03.06.25323548; doi: https://doi.org/10.1101/2025.03.06.25323548 Share This Article: Copy Citation Tools An AI-driven machine learning approach identifies risk factors associated with 30-day mortality following total aortic arch replacement combined with stent elephant implantation Shuai Zhang , Lulu Li , Jingyu Wang , Yuan Li , Cuntao Yu , Xiaogang Sun , Jing Sun , Xiangyang Qian medRxiv 2025.03.06.25323548; doi: https://doi.org/10.1101/2025.03.06.25323548 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Surgery Subject Areas All Articles Addiction Medicine (570) Allergy and Immunology (864) Anesthesia (302) Cardiovascular Medicine (4445) Dentistry and Oral Medicine (444) Dermatology (383) Emergency Medicine (609) Endocrinology (including Diabetes Mellitus and Metabolic Disease) (1515) Epidemiology (15236) Forensic Medicine (30) Gastroenterology (1127) Genetic and Genomic Medicine (6610) Geriatric Medicine (669) Health Economics (1000) Health Informatics (4549) Health Policy (1370) Health Systems and Quality Improvement (1613) Hematology (543) HIV/AIDS (1266) Infectious Diseases (except HIV/AIDS) (15926) Intensive Care and Critical Care Medicine (1104) Medical Education (623) Medical Ethics (147) Nephrology (668) Neurology (6613) Nursing (346) Nutrition (999) Obstetrics and Gynecology (1147) Occupational and Environmental Health (957) Oncology (3341) Ophthalmology (975) Orthopedics (369) Otolaryngology (420) Pain Medicine (436) Palliative Medicine (130) Pathology (665) Pediatrics (1694) Pharmacology and Therapeutics (693) Primary Care Research (714) Psychiatry and Clinical Psychology (5458) Public and Global Health (9244) Radiology and Imaging (2205) Rehabilitation Medicine and Physical Therapy (1370) Respiratory Medicine (1197) Rheumatology (596) Sexual and Reproductive Health (715) Sports Medicine (530) Surgery (713) Toxicology (99) Transplantation (289) Urology (265) (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'a0257a308960b910',t:'MTc3OTg4OTcwNw=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();