Editor's note
On April 6~9, 2024, Beijing time, the 73rd Annual Meeting of the American College of Cardiology (ACC.24) was held in Atlanta, USA. As the world's most influential highest-level conference in the field of cardiovascular disease, ACC has brought the latest progress and achievements of many scientific research to global cardiac experts and scholars. According to the latest data, there are more than 1 million patients with sudden acute myocardial infarction (AMI) in mainland China every year, and it is the leading cause of death in the world.
With the rapid development of social economy and the change of lifestyle of young people, the incidence of AMI is younger, which has become a major threat to public health and a social burden in contemporary society. The diagnosis, treatment, prediction and prevention of acute myocardial infarction (PMI) in young people are the focus of clinical attention. In recent years, the team of Professor Gao Jing of Tianjin Chest Hospital has done a large number of studies and reports on AMI in the ≤ of 45-year-old young people, and the team has summarized and analyzed the PMI data of a single center in the past 10 years, and released a number of latest original scientific research results on young myocardial infarction in the "Ischemic Heart Disease Session" and "Heart Failure and Cardiomyopathy Session" of ACC.24, showing the Voice of ACC China and bringing new inspiration to clinical diagnosis and treatment!
Study 1:
April 7 Beijing Time:
02：45-03：30
Advanced machine learning predicts coronary artery disease severity in patients with premature myocardial infarction
Study Two:
April 7 Beijing Time:
02：45-03：30
To develop a comprehensive algorithm for early myocardial infarction risk assessment in patients with heart failure with preserved ejection fraction (HFpEF) in hospitals
Study Three:
April 7 Beijing Time:
21：15-22：00
Circadian pattern of symptom onset and infarct size in patients with AMI under 45 years of age
Research 1
Session Title: Ischemic Heart Disease(缺血性心脏疾病专场)
ADVANCED MACHINE LEARNING FOR PREDICTING CORONARY ARTERY DISEASE SEVERITY IN PATIENTS WITH PREMATURE MYOCARDIAL INFARCTION
Advanced machine learning predicts coronary artery disease severity in patients with premature myocardial infarction
Jing Gao, Changping Li, Yuhang Wang, Jingxian Wang, Zhuang Cui, Yu Zhou, Anran Jing, Miaomiao Liang, Yin Liu, Tianjin Chest Hospital, Tianjin, People's Republic of China, Tianjin Medical University, Tianjin, People's Republic of China
Research Methods:
The study was a prospective, single-center, observational study, and consecutively included 1111 patients with AMI aged ≤45 years from January 2015 to January 2023 in Tianjin Chest Hospital. Lesion severity was divided into low-risk group (SYNTAX ≤22 points) and intermediate-high risk group (SYNTAX >22 points) according to the SYNTAX score on coronary angiography, and randomly assigned to the training or validation dataset in a 7:3 ratio. The influencing factors affecting the severity of coronary artery lesions in PMI patients were preliminarily screened out by Lasso-logistic regression, and the prediction model was established by using four machine learning methods, namely random forest (RF), K-nearest neighbor (KNN), support vector machine (SVM) and XGBoost, and the prediction system of coronary artery severity in PMI patients was established by validating the set data.
Findings:
Lasso-logistic regression identified seven important influencing factors: glycosylated hemoglobin, angina, apolipoprotein B, total bile acids, B-type brain natriuretic peptide (BNP), D-dimer, and fibrinogen. In the validation set, the area under the curve (AUC) of the prediction models established by random forest (RF), K-nearest neighbor (KNN), SVM and XGBoost were 0.775, 0.739, 0.656 and 0.800, the XGBoost model was the best prediction model (Fig. 1), and the model itself performed well in the validation, and finally an online prediction system for visualizing the severity of coronary artery lesions in PMI patients was established based on the XGBoost model, and the website is https://pmisyntax.shinyapps.io/appdecision/
Conclusions of the study
Based on machine learning, this study developed a novel and convenient visual risk prediction system that is clinically accessible and easy to evaluate. This makes the model expected to be a tool for early differentiation of lesion severity in PMI patients, which can quickly identify high-risk groups, optimize the current diagnosis and treatment system, and provide new tools and ideas for clinical decision-making.
Figure 1. A-LASSO regression lambda optimal penalty estimate, B-LASSO regression coefficient curve of predictor, C-ROC curve of different machine learning algorithms based on validation set, and relative importance of predictor of D-XGBoost.
Research 2
Session Title: Heart Failure and Cardiomyopathies (心力衰竭和心肌病专场)
DEVELOPING INTEGRATED ALGORITHMS FOR RISK ASSESSMENT OF IN-HOSPITAL HEART FAILURE WITH PRESERVED EJECTION FRACTION AMONG PATIENTS WITH PREMATURE MYOCARDIAL INFARCTION
To develop a comprehensive algorithm for early myocardial infarction risk assessment in patients with heart failure with preserved ejection fraction (HFpEF) in hospitals
Jing Gao, Jingxian Wang, Changping Li, Zhuang Cui, Yuhang Wang, Yu Zhou, Anran Jing, Miaomiao Liang, Yan Liang, Yin Liu, Tianjin Chest Hospital, Tianjin, People's Republic of China, Tianjin Medical University, Tianjin, People's Republic of China
Research Methods:
The study was a prospective, single-center, observational study that included 840 patients with AMI ≤ 45-year-old in Tianjin Chest Hospital. Lasso-Logistic, XGBoost, RF and KNN models were constructed to identify the risk of HFpEF during hospitalization in PMI patients.
Findings:
268 cases (31.9%) developed in-hospital HFpEF, and 572 cases (68.1%) did not develop heart failure. Patients were randomly assigned to the training set and the test set in an 8:2 ratio. The results show that the XGBoost model shows good performance in both the training set and the test set data (AUC: 0.848, Brier: 0.148). The final model included 10 predictors, which were BNP, SYNTAX score, age, monocyte-lymphocyte ratio, hematocrit, C-reactive protein-lymphocyte ratio, admission heart rate, body mass index (BMI), fibrinogen, and hypertension (Figure 2). Finally, a visual risk assessment and prediction system was developed https://hfpefpmi.shinyapps.io/apppredict/ to identify PMI patients with HFpEF in the hospital.
Conclusions of the study
The visual online prediction system shows good classification performance in identifying the risk of HFpEF in PMI patients hospitalized, which is convenient for clinical application, and is helpful to identify PMI high-risk groups with HFpEF risk in hospitals more accurately, so as to give more active treatment methods, improve prognosis, and improve the level of clinical diagnosis and treatment.
Figure 2. A-ROC curves of different models in the validation set, B-radar charts of performance comparison of different models, including precision, accuracy, AUC, F1 score and Brier score, importance ranking chart of predictors of C-XGBoost model, forest plot of predictors of D-XGBoost model.
Research 3
Session Title: Ischemic Heart Disease(缺血性心脏疾病专场)
CIRCADIAN PATTERN OF SYMPTOM ONSET AND INFARCT SIZE IN PATIENTS ≤45 YEARS WITH ACUTE MYOCARDIAL INFARCTION
Circadian pattern of symptom onset and infarct size in patients with AMI under 45 years of age
Jing Gao, Jing Ma, Yan Cui, Jingxian Wang, Wenqing Li, Yuhang Wang, Anran Jing, Miaomiao Liang, Yin Liu, Tianjin Chest Hospital, Tianjin, People's Republic of China, Tianjin Medical university, Tianjin, People's Republic of China
Research Methods:
A total of 1023 patients with ≤ 45-year-old AMI who had a record of onset time in Tianjin Chest Hospital were continuously enrolled. According to the time of symptom onset, they were divided into 4 groups at 6-hour intervals: 00:00-05:59, 06:00-11:59, 12:00-17:59 and 18:00-23:59. Periodic sinusoidal regression analysis was used to assess the circadian rhythm of infarct size (peak creatine kinase [CK] and peak creatine kinase isoenzyme [CKMB]). Multivariate linear regression analysis was used to analyze the effect of time to onset on infarct size.
Findings:
The peak incidence of AMI in young adults was 06:00-11:59 (P<0.05). Periodic sinusoidal regression analysis showed that the infarct area showed a significant circadian rhythm change with the 24-hour onset time (P<0.05), and the maximum infarct area appeared at 21 o'clock onset. Compared with the 06:00-11:59 group, the median peak CK and peak CKMB levels increased by 56.6% and 32.2%, respectively (P<0.0001). In the PPCI-treated youth acute ST-elevation myocardial infarction (STEMI) group, after adjusting for confounders by a multivariate linear regression model, the peak CK in the 18:00-23:59 group increased by an average of 660.7 U/L (95% CI: 165.2-1156.2) and the peak CKMB increased by an average of 68.0 U/L (95% CI: 33.6-102.4) compared with the 06:00-11:59 group (Fig. 3).
Conclusions of the study
This study found that the onset time of ≤ 45-year-old AMI patients had the circadian rhythm characteristics of "morning peak", and the infarct size showed a circadian rhythm with the 24-hour onset time. Time to onset should be a key factor in the evaluation of myocardial injury in young Chinese patients with AMI.
Figure 3. 1A-C distribution by onset time (6-hour interval), 2 A-C infarct area (peak CK) with onset time of 24 h, 3A-C infarct area (peak CKMB) with onset time of 24 h, 4A-B difference of peak CK and peak CKMB levels in different onset time groups, 5A-B multivariate linear regression analysis of STEMI onset time and infarct area of young people treated with PPCI
Expert Profile
Prof. Takashizu
Tianjin Chest Hospital
(Swipe up and down to view)
Ph.D., Postdoctoral Fellow, Chief Physician, Researcher
Deputy Director of the Institute of Cardiovascular Diseases, Tianjin Chest Hospital, Director of Biobank, Doctoral Supervisor of Tianjin University and Tianjin Medical University, and Postdoctoral Supervisor
Candidate of the first level of Tianjin "131" innovative talent training project, leader of Tianjin "131" innovative talent team, and the first high-level talent in Tianjin's health and family planning industry "Tianjin Medical Talent"
"The Most Beautiful Science and Technology Worker" in Tianjin Health System
Expert Committee Member of the American College of Cardiology (FACC), Expert Member of the European Society of Cardiology (FESC), Professional Member of the American College of Cardiology (AHA), Expert of the Review Committee of the European Cardiology Annual Meeting.
Expert of the National Science and Technology Program Expert Database, Expert Service Key Contact Expert of the Ministry of Human Resources and Social Security of the People's Republic of China
High-level talent evaluation expert of the Talent Program of the Ministry of Education of the People's Republic of China
Evaluation expert of the Academic Degrees and Graduate Education Development Center of the Ministry of Education of the People's Republic of China
Evaluation expert of China Postdoctoral Science Foundation
Expert of the Science and Technology Expert Database of Shanghai and Tianjin Science and Technology Bureau
Deputy head of the Beijing-Tianjin-Hebei Biological Sample Resources Collaboration Group of the Clinical Data and Sample Resource Library Committee of the Chinese Association of Research Hospitals
Vice Chairman of the Beijing-Tianjin-Hebei Biobank Alliance
Vice Chairman of the Basic Research Committee of the Cardiovascular Physician Branch of Tianjin Medical Doctor Association
Director of Tianjin Cardiology Society
Reviewers or editorial board members of more than 10 SCI indexed journals
He has presided over and participated in 27 scientific research projects at all levels, such as national, provincial and ministerial key research and development, major projects, and key discipline projects
His representative achievements have been presented and exchanged at the world's top international conferences in the field of cardiovascular ACC, ESC, AHA, TCT for nearly 22 times. It has won 13 municipal scientific and technological achievements and 7 national patents. He has won 2 second prizes of Tianjin Science and Technology Progress Award, published more than 100 papers, and trained more than 55 doctoral and master's students.
Jing Ma is an assistant researcher
Tianjin Chest Hospital
Postdoctoral fellow at the station, assistant researcher. Department of Epidemiology and Health Statistics, research interests: epidemiology and basic research of early-onset acute myocardial infarction. He is a core member of Tianjin's "131" innovative talent team. In the past 3 years, he has published 1 SCI paper as a co-first author, participated in the publication of 5 SCI papers, and 7 Chinese core papers. The first author completed 1 abstract included in the 2024 American College of Cardiology Annual Meeting (ACC), and the first/co-first author completed 2 abstracts included in the 2021 European Society of Cardiology Annual Meeting (ESC). He presided over 1 scientific and technological talent cultivation project of Tianjin Municipal Health Commission, and the first person to complete it was awarded 1 Tianjin scientific and technological achievement, participated in national key research and development, Tianjin major projects, key scientific and technological support projects of key R&D plans and key disciplines of Tianjin health science and technology projects, and participated in the completion of 6 scientific and technological achievements of Tianjin.
Dr. Wang Jingxian
Tianjin Chest Hospital
Ph.D. candidate, research interests: machine learning algorithm modeling, multi-omics, lipid metabolism and acute myocardial infarction clinical research and basic research. During the degree study, he completed 1 abstract of the 2024 American Cardiology Annual Meeting (ACC) as the first author, participated in the completion of 5 abstracts of the 2023 and 2024 American Cardiology Annual Meeting (ACC), 2 abstracts included in the European Society of Cardiology Annual Meeting (ESC), and published 1 SCI paper as the first author. He won the third prize of the graduate group of the Tianjin Division of the 2023 (9th) National College Student Statistical Modeling Competition. Participated in the key R&D of the Ministry of Science and Technology of the People's Republic of China, the major special projects of the Tianjin Municipal Science and Technology Commission, the key scientific and technological support key projects of the key R&D plan, and the key discipline special projects of the Tianjin Health Science and Technology Project.
Dr. Wang Yuhang
Tianjin Chest Hospital
Master's degree candidate, research direction: machine learning and the establishment of risk prediction models for early-onset acute myocardial infarction. During his degree studies, he participated in the publication of 1 SCI article, co-first author published 1 abstract of the 2024 American Cardiology Annual Meeting (ACC), participated in the completion of 3 abstracts of the 2024 American Cardiology Annual Meeting (ACC), and won the third prize of the graduate group of the Tianjin Division of the 2023 (9th) National College Student Statistical Modeling Competition. Participated in the "13th Five-Year Plan" Key R&D Program of the Ministry of Science and Technology of the People's Republic of China, the Key Research and Development Project of Science and Technology Support of Tianjin Key R&D Program in 2020, and the Key Specialty Project of Tianjin Municipal Health Commission.