Establishment of a diagnostic model of coronary heart disease in elderly patients with diabetes mellitus based on machine learning algorithms.

OBJECTIVE: To establish a prediction model of coronary heart disease (CHD) in elderly patients with diabetes mellitus (DM) based on machine learning (ML) algorithms. METHODS: Based on the Medical Big Data Research Centre of Chinese PLA General Hospital in Beijing, China, we identified a cohort of elderly inpatients (≥ 60 years), including 10,533 patients with DM complicated with CHD and 12,634 patients with DM without CHD, from January 2008 to December 2017. We collected demographic characteristics and clinical data. After selecting the important features, we established five ML models, including extreme gradient boosting (XGBoost), random forest (RF), decision tree (DT), adaptive boosting (Adaboost) and logistic regression (LR). We compared the receiver operating characteristic curves, area under the curve (AUC) and other relevant parameters of different models and determined the optimal classification model. The model was then applied to 7447 elderly patients with DM admitted from January 2018 to December 2019 to further validate the performance of the model. RESULTS: Fifteen features were selected and included in the ML model. The classification precision in the test set of the XGBoost, RF, DT, Adaboost and LR models was 0.778, 0.789, 0.753, 0.750 and 0.689, respectively; and the AUCs of the subjects were 0.851, 0.845, 0.823, 0.833 and 0.731, respectively. Applying the XGBoost model with optimal performance to a newly recruited dataset for validation, the diagnostic sensitivity, specificity, precision, and AUC were 0.792, 0.808, 0.748 and 0.880, respectively. CONCLUSIONS: The XGBoost model established in the present study had certain predictive value for elderly patients with DM complicated with CHD.

Lipoprotein(a) is associated with coronary atheroma progression: analysis from a serial coronary computed tomography angiography study.

BACKGROUND: Lipoprotein(a) [Lp(a)] has been closely related to coronary atherosclerosis and might affect perivascular inflammation due to its proinflammatory properties. However, there are limited data about Lp(a) and related perivascular inflammation on coronary atheroma progression. Therefore, this study aimed to investigate the associations between Lp(a) and the perivascular fat attenuation index (FAI) with coronary atheroma progression detected by coronary computed tomography angiography (CCTA). METHODS: Patients who underwent serial CCTA examinations without a history of revascularization and with available data for Lp(a) within one month before or after baseline and follow-up CCTA imaging scans were considered to be included. CCTA quantitative analyses were performed to obtain the total plaque volume (TPV) and the perivascular FAI. Coronary plaque progression (PP) was defined as a ≥ 10% increase in the change of the TPV at the patient level or the presence of new-onset coronary atheroma lesions. The associations between Lp(a) or the perivascular FAI with PP were examined by multivariate logistic regression. RESULTS: A total of 116 patients were ultimately enrolled in the present study with a mean CCTA interscan interval of 30.80 ± 13.50 months. Among the 116 patients (mean age: 53.49 ± 10.21 years, males: 83.6%), 32 patients presented PP during the follow-up interval. Lp(a) levels were significantly higher among PP patients than those among non-PP patients at both baseline [15.80 (9.09-33.60) mg/dLvs. 10.50 (4.75-19.71) mg/dL,P = 0.029] and follow-up [20.60 (10.45-34.55) mg/dLvs. 8.77 (5.00-18.78) mg/dL,P = 0.004]. However, there were no differences in the perivascular FAI between PP group and non-PP group at either baseline or follow-up. Multivariate logistic regression analysis showed that elevated baseline Lp(a) level (OR = 1.031, 95% CI: 1.005-1.058,P = 0.019) was an independent risk factor for PP after adjustment for other conventional variables. CONCLUSIONS: Lp(a) was independently associated with coronary atheroma progression beyond low-density lipoprotein cholesterol and other conventional risk factors. Further studies are warranted to identify the inflammation effect exhibited as the perivascular FAI on coronary atheroma progression.

Prediction of fatal adverse prognosis in patients with fever-related diseases based on machine learning: A retrospective study.

BACKGROUND: Fever is the most common chief complaint of emergency patients. Early identification of patients at an increasing risk of death may avert adverse outcomes. The aim of this study was to establish an early prediction model of fatal adverse prognosis of fever patients by extracting key indicators using big data technology. METHODS: A retrospective study of patients' data was conducted using the Emergency Rescue Database of Chinese People's Liberation Army General Hospital. Patients were divided into the fatal adverse prognosis group and the good prognosis group. The commonly used clinical indicators were compared. Recursive feature elimination (RFE) method was used to determine the optimal number of the included variables. In the training model, logistic regression, random forest, adaboost and bagging were selected. We also collected the emergency room data from December 2018 to December 2019 with the same inclusion and exclusion criterion. The performance of the model was evaluated by accuracy, F1-score, precision, sensitivity and the areas under receiver operator characteristic curves (ROC-AUC). RESULTS: The accuracy of logistic regression, decision tree, adaboost and bagging was 0.951, 0.928, 0.924, and 0.924, F1-scores were 0.938, 0.933, 0.930, and 0.930, the precision was 0.943, 0.938, 0.937, and 0.937, ROC-AUC were 0.808, 0.738, 0.736, and 0.885, respectively. ROC-AUC of ten-fold cross-validation in logistic and bagging models were 0.80 and 0.87, respectively. The top six coefficients and odds ratio (OR) values of the variables in the Logistic regression were cardiac troponin T (CTnT) (coefficient=0.346, OR = 1.413), temperature (T) (coefficient=0.235, OR = 1.265), respiratory rate (RR) (coefficient= -0.206,OR = 0.814), serum kalium (K) (coefficient=0.137, OR = 1.146), pulse oxygen saturation (SPO2) (coefficient= -0.101, OR = 0.904), and albumin (ALB) (coefficient= -0.043, OR = 0.958). The weights of the top six variables in the bagging model were: CTnT, RR, lactate dehydrogenase, serum amylase, heartrate, and systolic blood pressure. CONCLUSIONS: The main clinical indicators of concern included CTnT, RR, SPO2, T, ALB and K. The bagging model and logistic regression model had better diagnostic performance comprehesively. Those may be conducive to the early identification of critical patients with fever by physicians.