Aim: Heart disease detection using machine learning methods has been an outstanding research topic as heart diseases continue to be a burden on healthcare systems around the world. Therefore, in this study, the performances of machine learning methods for predictive classification of coronary heart disease were compared.Material and Method: In the study, three different models were created with Random Forest (RF), Logistic Regression (LR), and Support Vector Machine (SVM) algorithms for the classification of coronary heart disease. For hyper parameter optimization, 3-repeats 10-fold repeated cross validation method was used. The performance of the models was evaluated based on Accuracy, F1 Score, Specificity, Sensitivity, Positive Predictive Value, Negative Predictive Value, and Confusion Matrix (Classification matrix).Results: RF 0.929, SVM 0.897 and LR 0.861 classified coronary heart disease with accuracy. Specificity, Sensitivity, F1-score, Negative predictive and Positive predictive values of the RF model were calculated as 0.929, 0.928, 0.928, 0.929 and 0.928, respectively. The Sensitivity value of the SVM model was higher compared to the RF. Conclusion: Considering the accurate classification rates of Coronary Heart disease, the RF model outperformed the SVM and LR models. Also, the RF model had the highest sensitivity value. We think that this result, which has a high sensitivity criterion in order to minimize overlooked heart patients, is clinically very important.

Background/Objectives: Acute decompensated heart failure (ADHF) presents a significant clinical challenge characterized by frequent hospitalizations, high mortality rates, and substantial healthcare costs. The united index of hemoglobin, albumin, lymphocytes and platelets (HALP) is a new indicator that reflects systemic inflammation and nutritional status. This study aimed to investigate the prognostic utility of the HALP score and hematological parameters in predicting short-term mortality among ADHF patients admitted to the coronary care unit (CCU). Methods: This investigation adopts a retrospective observational design, encompassing a cohort of patients with ADHF who were followed in the CCU at our medical institution between January 2019 and April 2024. Results: The cohort of 227 individuals was dichotomized into two subsets based on the presence or absence of short-term mortality in the hospital, resulting in 163 (71.8%) and 64 (28.2%) individuals in the survivor and exitus groups, respectively. Age was significantly higher in the exitus group (p-value = 0.004). Hemoglobin, lymphocyte count, platelet count, albumin, and HALP score were significantly higher in the survivor group (all p-values < 0.001). No significant difference was observed between the groups in terms of gender, diabetes mellitus (DM), coronary artery disease (CAD), or ejection fraction (EF), although hypertension (HT) prevalence was significantly higher in the exitus group (p-value = 0.038). ROC analysis demonstrated that hemoglobin, lymphocyte, albumin, and HALP score had significant discriminative power, with albumin showing the highest AUC (0.814). Conclusions: In conclusion, the HALP score and hematological parameters represent valuable prognostic feature for short-term mortality prediction in ADHF patients admitted to the CCU. These findings underscore the importance of early risk stratification and targeted interventions guided by comprehensive biomarker assessments in optimizing patient outcomes.

Background: The Uric acid/Albumin ratio (UAR) has recently been identified as a prominent marker in cardiovascular diseases. In this study, we aimed to reveal the effect of UAR on coronary collateral circulation (CCC) in patients with stable coronary artery disease (CAD) patients by comparing it with conventional inflammation-based markers. Methods: In this study, 415 consecutive patients who underwent coronary angiography for stable angina pectoris and were found to have chronic total occlusion in at least one coronary artery were retrospectively included. The study population was divided into two groups as good CCC (Rentrop 2–3) and poor CCC (Rentrop 0–1) according to the Rentrop classification, and the groups were compared in terms of UAR and other traditional inflammation-based markers. Results: In the poor CCC group, C-reactive protein/albumin ratio (CAR), monocyte/high-density lipoprotein cholesterol ratio (MHR), neutrophil/lymphocyte ratio (NLR), platelet/lymphocyte ratio (PLR), systemic immune-inflammation index (SII) and UAR were found to be significantly high ( p < .05, for all). UAR negatively correlated with rentrop classification (r = −0.383, p < .001). In multivariate regression analysis, MHR, NLR, SII and UAR were determined as independent predictors for poor CCC ( p < .05, for all). The ability of UAR to predict poor CCC was superior to uric acid and albumin alone ( p < .0001, for both). In addition, UAR was found to be superior to other inflammation-based markers in predicting poor CCC ( p < .005, for all). Conclusion: UAR was identified as a strong and independent predictor of CCC. In this context, UAR may be a useful biomarker in the risk prediction of patients with stable CAD.

Acute myocardial infarction (AMI) is the main cause of death in developed and developing countries. AMI is a serious medical problem that necessitates hospitalization and sometimes results in death. Patients hospitalized in the emergency department (ED) should therefore receive an immediate diagnosis and treatment. Many studies have been conducted on the prognosis of AMI with hemogram parameters. However, no study has investigated potential hemogram parameters for the diagnosis of AMI using an interpretable artificial intelligence-based clinical approach. The purpose of this research is to implement the principles of explainable artificial intelligence (XAI) in the analysis of hematological predictors for AMI. In this retrospective analysis, 477 (48.6%) patients with AMI and 504 (51.4%) healthy individuals were enrolled and assessed in predicting AMI. Of the patients with AMI, 182 (38%) had an ST-segment elevation MI (STEMI), and 295 (62%) had a non-ST-segment elevation MI (NSTEMI). Demographic and hematological information of the patients was analyzed to determine AMI. The XAI approach combined with machine learning approaches (Extreme Gradient Boosting, XGB; Adaptive Boosting, AB; Light Gradient Boosting Machine, LGBM) was applied for the estimation of AMI and distinguishing subgroups of AMI (STEMI and NSTEMI). The SHAP approach was used to explain the predictions intuitively. After selecting the 10 most important hematological parameters for AMI, the LGBM model achieved 83% and 74% accuracy for prediction of AMI, and distinguishing subgroups of AMI (STEMI and NSTEMI), respectively. SHAP results showed that neutrophil (NEU), white blood cell (WBC), platelet width of distribution (PDW), and basophil (BA) were the most important for AMI prediction. Mean corpuscular volume (MCV), BA, monocytes (MO), and lymphocytes (LY) were the most important hematological parameters that distinguish STEMI from NSTEMI. The proposed model serves as a valuable tool for physicians, facilitating the diagnosis, treatment, and follow-up of patients with AMI and distinguishing subgroups of AMI (STEMI and NSTEMI). Analyzing readily accessible hemogram parameters empowers medical professionals to make informed decisions and provide enhanced care to a wide range of individuals.

— Aim: The aim of this study is to develop a predictive classification model that can identify risk factors for heart attack disease. Materials and Methods: In the study, patients with low and high probability of having a heart attack were examined. Variable importance was calculated to identify risk factors. The radial basis function and multilayer perception neural networks were used to compare the classification prediction results. Results: MLP model criteria; Accuracy 0.911, F1 score 0.918, Specificity 0.92, Sensitivity 0.903, while RBF model criteria were obtained as accuracy 0.797, F1 score 0.812, Specificity 0.84, Sensitivity 0.765. The first three most important factors that may be associated with having a heart attack were obtained as trestbps, oldpeak, and chol. Conclusion: According to the prediction results of the heart attack, it can be said that the model created with the MLP neural network has more successful predictions than the model created with the RBF neural network. In addition, estimating the importance values of the factors most associated with heart attack (obtaining the most important biomarkers that may cause heart attack) is a promising result for the diagnosis, treatment and prognosis of the disease. Keywords— Heart Attack, machine learning, neural networks, classification, variable importance.

Background and Objectives: Acute coronary syndrome (ACS) is a life-threatening cardiovascular condition with high mortality rates, necessitating accurate and early risk assessment to optimize patient outcomes. While traditional lipid markers, such as low-density lipoprotein-cholesterol (LDL-C) and high-density lipoprotein-cholesterol (HDL-C), are widely used, non-traditional lipid indices, including the lipoprotein combined index (LCI), atherogenic index of plasma (AIP), atherogenic index (AI), Castelli risk indices (CRI-I, CRI-II), and atherogenic combined index (ACI) may offer additional prognostic insights by reflecting the underlying atherogenic and inflammatory processes. This study aimed to assess the prognostic value of these non-traditional lipid indices, along with traditional lipid and biochemical markers, for in-hospital mortality in ACS patients. Materials and Methods: This retrospective observational study analyzed data from ACS patients admitted to the coronary care unit (CCU) between January 2019 and September 2024. A cohort of 920 patients was divided into survivor (n = 823, 89.46%) and non-survivor (n = 97, 10.54%) groups based on in-hospital mortality outcomes. Demographic, hematological, biochemical, and lipid profile data, including traditional and non-traditional lipid indices, were collected. Separate logistic regression models were developed for each index, adjusting for demographic and clinical variables in order to assess the independent predictive power of each non-traditional lipid index. Results: Significant differences were observed between survivor and non-survivor groups in terms of age, c-reactive protein (CRP), white blood cell count (WBC), hemoglobin (HGB), and creatinine levels (all p-values < 0.05). While traditional lipid markers, such as LDL-C and HDL-C, showed limited predictive value, non-traditional lipid indices demonstrated stronger associations. The highest Exp (Beta) values were observed for the CRI-II, AI, and CRI-I. An ROC analysis further confirmed that the CRI-II, AI, and CRI-I had the highest AUC values, with pairwise comparisons underscoring the CRI-II’s superior accuracy. These findings suggest that non-traditional lipid indices predict atherogenic risk better than traditional markers alone. Conclusions: Non-traditional lipid indices, particularly the CRI-I and II, AI, LCI, ACI, and AIP, were found to be significantly associated with in-hospital mortality in ACS patients. These indices may provide additional prognostic value beyond traditional lipid parameters; however, further prospective studies are needed to confirm their clinical utility. These results underscore the importance of integrating non-traditional lipid indices into routine risk assessments to improve mortality predictions and inform targeted interventions in high-risk ACS patients.