Explainable and uncertainty-aware ensemble framework with causal analysis for breast cancer detection.

Background To assess the feasibility of a machine learning (ML) approach using radiomics features of perfusion defects on rest myocardial perfusion imaging (MPI) to detect the presence of hibernating myocardium. Methodology Data of patients who underwent 99mTc-sestamibi MPI and 18F-FDG PET/CT for myocardial viability assessment were retrieved. Rest MPI data were processed on ECToolbox, and polar maps were saved using theNFile PMap tool. The reference standard for defining hibernating myocardium was the presence of mismatched perfusion-metabolism defect with impaired myocardial contractility at rest. Perfusion defects on the polar maps were delineated with regions of interest (ROIs) after spatial resampling and intensity discretization. Replicable random sampling allocated 80% (257) of the perfusion defects of the patients from January 2017 to September 2022 to the training set and the remaining 20% (64) to the validation set. An independent dataset of perfusion defects from 29 consecutive patients from October 2022 to January 2023 was used as the testing set for model evaluation. One hundred ten first and second-order texture features were extracted for each ROI. After feature normalization and imputation, 14 best-ranked features were selected using a multistep feature selection process including the Logistic Regression and Fast Correlation-Based Filter. Thirteen supervised ML algorithms were trained with stratified five-fold cross-validation on the training set and validated on the validation set. The ML algorithms with a Log Loss of <0.688 and <0.672 in the cross-validation and validation steps were evaluated on the testing set. Performance matrices of the algorithms assessed included area under the curve (AUC), classification accuracy (CA), F1 score, precision, recall, and specificity. To provide transparency and interpretability, SHapley Additive exPlanations (SHAP) values were assessed and depicted as beeswarm plots. Results Two hundred thirty-nine patients (214 males; mean age 56 ± 11 years) were enrolled in the study. There were 371 perfusion defects (321 in the training and validation sets; 50 in the testing set). Based on the reference standard, 168 perfusion defects had hibernating myocardium (139 in the training and validation sets; 29 in the testing set). On cross-validation, six ML algorithms with Log Loss <0.688 had AUC >0.800. On validation, 10 ML algorithms had a Log Loss value <0.672, among which six had AUC >0.800. On model evaluation of the selected models on the unseen testing set, nine ML models had AUC >0.800 with Gradient Boosting Random Forest (xgboost) [GB RF (xgboost)] achieving the highest AUC of 0.860 and could detect the presence of hibernating myocardium in 21/29 (72.4%) perfusion defects with a precision of 87.5% (21/24), specificity 85.7% (18/21), CA 78.0% (39/50) and F1 Score 0.792. Four models depicted a clear pattern of model interpretability based on the beeswarm SHAP plots. These were GB RF (xgboost), GB (scikit-learn), GB (xgboost), and Random Forest. Conclusion Our study demonstrates the potential of ML in detecting hibernating myocardium using radiomics features extracted from perfusion defects on rest MPI images. This proof-of-concept underscores the notion that radiomics features capture nuanced information beyond what is perceptible to the human eye, offering promising avenues for improved myocardial viability assessment.

&lt;p&gt;Breast cancer is ranked as a significant cause of mortality among females globally. Its complex nature poses principal challenges for physicians and researchers for rapid diagnosis and prognosis. Hence, machine learning algorithms are employed to forecast and identify diseases. This study discusses the comparative analysis of seven machine learning models, e.g., logistic regression (LR), support vector machine (SVM), k-nearest neighbor classifier (KNN), decision tree classifier (DT), random forest classifier (RF), Naïve Bayes (NB), and artificial neural network (ANN) to predict breast cancer using Wisconsin breast cancer and breast cancer datasets. In the Wisconsin breast cancer dataset, KNN depicted 99% accuracy, followed by RF (98%), SVM (96%), NB (96%), LR (96%), ANN (93%), and DT (92%). On the contrary, in the breast cancer (BC) dataset, the highest accuracy was achieved by LR at 83%, and the lowest was achieved by DT (65%), which depicted that the numeric dataset WBC has better accuracy than the breast cancer dataset.&lt;/p&gt;

Prediction of gully erosion susceptibility through the lens of the SHapley Additive exPlanations (SHAP) method using a stacking ensemble model.

Background Predicting the prognosis of patients with acute myocardial infarction (AMI) combined with diabetes mellitus (DM) is crucial due to high in-hospital mortality rates. This study aims to develop and validate a mortality risk prediction model for these patients by interpretable machine learning (ML) methods. Methods Data were sourced from the Medical Information Mart for Intensive Care IV (MIMIC-IV, version 2.2). Predictors were selected by Least absolute shrinkage and selection operator (LASSO) regression and checked for multicollinearity with Spearman’s correlation. Patients were randomly assigned to training and validation sets in an 8:2 ratio. Seven ML algorithms were used to construct models in the training set. Model performance was evaluated in the validation set using metrics such as area under the curve (AUC) with 95% confidence interval (CI), calibration curves, precision, recall, F1 score, accuracy, negative predictive value (NPV), and positive predictive value (PPV). The significance of differences in predictive performance among models was assessed utilising the permutation test, and 10-fold cross-validation further validated the model’s performance. SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME) were applied to interpret the models. Results The study included 2,828 patients with AMI combined with DM. Nineteen predictors were identified through LASSO regression and Spearman’s correlation. The Random Forest (RF) model was demonstrated the best performance, with an AUC of 0.823 (95% CI: 0.774–0.872), high precision (0.867), accuracy (0.873), and PPV (0.867). The RF model showed significant differences (p < 0.05) compared to the K-Nearest Neighbours and Decision Tree models. Calibration curves indicated that the RF model’s predicted risk aligned well with actual outcomes. 10-fold cross-validation confirmed the superior performance of RF model, with an average AUC of 0.828 (95% CI: 0.800–0.842). Significant Variables in RF model indicated that the top eight significant predictors were urine output, maximum anion gap, maximum urea nitrogen, age, minimum pH, maximum international normalised ratio (INR), mean respiratory rate, and mean systolic blood pressure. Conclusion This study demonstrates the potential of ML methods, particularly the RF model, in predicting in-hospital mortality risk for AMI patients with DM. The SHAP and LIME methods enhance the interpretability of ML models.

ABSTRACTBackgroundAcute myocardial infarction (AMI) remains a leading global cause of mortality. This study explores predictors of in‐hospital mortality among AMI patients using advanced machine learning (ML) techniques.MethodsData from 7422 AMI patients treated with percutaneous coronary intervention (PCI) at Tehran Heart Center (2015–2021) were analyzed. Fifty‐eight clinical, demographic, and laboratory variables were evaluated. Seven ML algorithms, including Random Forest (RF), logistic regression with LASSO, and XGBoost, were implemented. The data set was divided into training (70%) and testing (30%) subsets, with fivefold cross‐validation. The class imbalance was addressed using the synthetic minority oversampling technique (SMOTE). Model predictions were interpreted using SHapley Additive exPlanations (SHAP).ResultsIn‐hospital mortality occurred in 129 patients (1.74%). RF achieved the highest predictive performance, with an area under the curve (AUC) of 0.924 (95% CI 0.893–0.954), followed by XGBoost (AUC 0.905) and logistic regression with LASSO (AUC 0.893). Sensitivity analysis in STEMI patients confirmed RF's robust performance (AUC 0.900). SHAP analysis identified key predictors, including lower left ventricular ejection fraction (LVEF; 33.24% vs. 43.46% in survivors, p < 0.001), higher fasting blood glucose (190.38 vs. 132.29 mg/dL, p < 0.001), elevated serum creatinine, advanced age (70.92 vs. 61.88 years, p < 0.001), and lower LDL‐C levels. Conversely, BMI showed no significant association (p = 0.456).ConclusionML algorithms, particularly RF, effectively predict in‐hospital mortality in AMI patients, highlighting critical predictors such as LVEF and biochemical markers. These insights offer valuable tools for enhancing clinical decision‐making and improving patient outcomes.

As the global population ages, cardiovascular diseases, particularly heart failure (HF), have become leading causes of mortality and disability among elderly patients. Diabetes and hypertension are major risk factors for cardiovascular diseases, making this group especially vulnerable to HF. Current clinical tools for predicting HF risk are often complex, requiring extensive clinical parameters and laboratory tests, which limit their practical application. Therefore, a need exists for a predictive model that is both simple and effective in assessing HF risk in elderly patients with diabetes and hypertension. This study utilized data from the National Health and Nutrition Examination Survey, spanning seven cycles from 2003 to 2016, including 71 058 subjects. The study focused on elderly patients (aged 65 and above) diagnosed with both diabetes and hypertension, ultimately including 1445 participants. We examined seven novel composite indices: a body shape index (ABSI), atherogenic index of plasma (AIP), BARD score, body fat percentage (BFP), body roundness index (BRI), fatty liver index (FLI), and prognostic nutritional index (PNI). These indices were selected for their simplicity and ease of calculation from routine clinical assessments. The primary outcome was HF status, and data pre-processing included imputation for missing values using random forest algorithms. Various machine learning models were applied, including random forest, logistic regression, XGBoost, and others, with model performance assessed through metrics like accuracy, precision, recall, F1 score, and receiver operating characteristic-area under the curve (ROC AUC). The best-performing model was further analysed using SHAP (SHapley Additive exPlanations) values to determine feature importance. The study found that the XGBoost model demonstrated superior performance across all evaluation metrics, with an area under the curve (AUC) value of 0.96. Significant predictors of HF included BRI and PNI, which had the highest SHAP values, indicating their substantial influence on model predictions. The study also highlighted the robust predictive capabilities of AIP, particularly in assessing cardiovascular events in elderly patients. The study demonstrates that novel composite indices like ABSI, AIP, BARD score, BFP, BRI, FLI, and PNI have significant potential in predicting HF risk among elderly diabetic and hypertensive patients. These indices offer clinicians new tools for cardiovascular risk assessment that are simpler and potentially more effective in clinical practice. Future research should focus on validating these findings in different populations and exploring their longitudinal predictive power. This study explores simple and effective ways to predict heart failure risk in elderly patients with diabetes and hypertension by using novel, easy-to-calculate indices.Indices like ABSI, AIP, BARD score, BFP, BRI, FLI, and PNI are highly effective in predicting heart failure risk in elderly diabetic and hypertensive patients.The XGBoost machine learning model, which uses these indices, demonstrated strong predictive power with an AUC of 0.96, highlighting its clinical applicability.

BackgroundHormone receptor (HR)-positive, human epidermal growth factor receptor 2 (HER2)-negative breast cancer is the most prevalent subtype among women but has a modest response to neoadjuvant chemotherapy (NAC). Accurately predicting NAC efficacy and recurrence risk remains challenging, as conventional clinical and molecular markers have limited predictive power. Advances in digital pathology and artificial intelligence now enable quantitative pathomics analysis, offering new opportunities for precise prediction and prognostic assessment.MethodsIn this retrospective study, 162 HR-positive, HER2-negative breast cancer patients treated with NAC between 2014 and 2021 were included. Hematoxylin and eosin (H&E)-stained pretreatment biopsy slides were digitized and analyzed using Vision Transformer (ViT) and Unified Network for Image (UNI) deep learning models to extract pathomic features. Thirteen clinical variables were collected. After least absolute shrinkage and selection operator (LASSO)-based feature selection, multiple machine learning models were developed for both response prediction and prognostic evaluation of recurrence, with performance evaluated by receiver operating characteristic (ROC) curves, area under the curve (AUC), sensitivity, specificity, confusion matrix, calibration curves, and decision curve analysis (DCA). Furthermore, SHapley Additive exPlanations (SHAP) was used to rank the importance of features for each model.ResultsThe CatBoost model achieved the best predictive performance (AUC = 0.900 in training and 0.848 in validation) when a combination of clinical and pathomics-derived variables was used. Key predictive factors included Ki-67 expression, age, histological grade, PR status, and prominent pathomic features. A Kaplan–Meier survival plot indicated that regardless of stratification by MP grade or pCR status, there was no significant difference in recurrence status or survival outcomes between the two groups in this cohort. Furthermore, the recurrence models developed mainly using pathomics were strongly accurate for predicting 1-year recurrence (AUC = 0.907 in training and 0.769 in validation).ConclusionsIntegrating pathomic features with clinical variables via machine learning enables robust pretreatment prediction of NAC efficacy and short-term recurrence in HR-positive, HER2-negative breast cancer. This approach has the potential to offer a clinically practical tool to optimize individualized therapy and improve patient management, highlighting the translational value of AI-powered digital pathology in breast cancer care.

As the most prevalent cancer in women worldwide, breast cancer requires early detection to have the best possible treatment results. However, conventional screening techniques, like mammography and clinical examinations, can take time and effort. In this paper, we proposed a predictive model for the identification of breast cancer by combining state-of-the-art model BERT with machine learning approaches. Several machine learning algorithms, such as K-nearest neighbours, decision tree, random forest, neural network, and BERT, were tested on the Breast Cancer Wisconsin (Diagnostic) Dataset for early prediction of the diseases. The BERT model's accuracy was improved by using hyperparameter optimisation techniques. For the proposed work's evaluation, we used metrics like accuracy, F1-score, precision, and recall on the standard publically available datasets. With an accuracy of 0.98 across various splits and an area under the curve (AUC) of 0.98 in receiver operating characteristic (ROC) curves, our results show that BERT consistently works better than other models. These findings highlight the value of early and reliable identification in improving patient outcomes, highlighting the promise of machine learning algorithms, notably BERT, inaccurate breast cancer prediction.

Chronic Hepatitis B (CHB) is a leading cause of liver fibrosis. Accurate and non-invasive diagnosis of liver fibrosis in CHB patients is of critical clinical importance. This study aimed to develop and validate machine learning (ML)-based models for predicting significant liver fibrosis in CHB patients. This retrospective cohort study included 328 CHB patients (225 with non-significant liver fibrosis and 103 with significant liver fibrosis) from 2017 to 2022. Four ML models were constructed based on four selected features identified through the least absolute shrinkage and selection operator (LASSO) regression. Model performance was assessed using the receiver operating characteristic (ROC) curve, and the area under the curve (AUC), accuracy, sensitivity, specificity, and SHapley Additive exPlanations (SHAP) analysis. The random forest (RF) model demonstrated the highest predictive performance, with an AUC of 0.874 (95% CI: 0.813-0.934) in the training set and 0.863 (95% CI: 0.772-0.955) in the test set, outperforming extreme gradient boosting (XGBoost), logistic regression (LR), and support vector machine (SVM). Compared with the traditional fibrosis indices such as aspartate aminotransferase to platelet ratio index (APRI) (AUC = 0.585) and fibrosis-4 (FIB-4) (AUC = 0.633), the RF model (AUC = 0.863) demonstrated significantly higher predictive accuracy. SHAP analysis identified platelet count (PLT) as the most influential predictor in the RF model. The ML-based RF model offers a highly accurate, non-invasive interpretable tool for predicting significant liver fibrosis in patients with CHB, offering potential for clinical application in routine fibrosis risk assessment.

BackgroundDepression is major global public health problems among university students. Currently, the evaluation and monitoring of depression predominantly depend on subjective and self-reported methods. There is an urgent necessity to develop objective means of identifying depression. Acoustic features, which convey emotional information, have the potential to enhance the objectivity of depression assessments. This study aimed to investigate the feasibility of utilizing acoustic features for the objective and automated identification and characterization of depression among Chinese university students.MethodsA cross-sectional study was undertaken involving 103 students with depression and 103 controls matched for age, gender, and education. Participants' voices were recorded using a smartphone as they read neutral texts. Acoustic analysis and feature extraction were performed using the OpenSMILE toolkit, yielding 523 features encompassing spectral, glottal, and prosodic characteristics. These extracted acoustic features were utilized for discriminant analysis between depression and control groups. Pearson correlation analyses were conducted to evaluate the relationship between acoustic features and Patient Health Questionnaire-9 (PHQ-9) scores. Five machine learning algorithms including Linear Discriminant Analysis (LDA), Logistic Regression, Support Vector Classification, Naive Bayes, and Random Forest were used to perform the classification. For training and testing, ten-fold cross-validation was employed. Model performance was assessed using receiver operating characteristic (ROC) curve, area under the curve (AUC), precision, accuracy, recall, and F1 score. Shapley Additive exPlanations (SHAP) method was used for model interpretation.ResultsIn depression group, 32 acoustic features (25 spectral features, 5 prosodic features and 2 glottal features) showed significant alterations compared with controls. Further, 27 acoustic features (10 spectral features, 3 prosodic features, and 1 glottal features) were significantly correlated with depression severity. Among five machine learning algorithms, LDA model demonstrated the highest classification performance, with an AUC of 0.771. SHAP analysis suggested that Mel-frequency cepstral coefficients (MFCC) features contributed most to the model's classification efficacy.ConclusionsThe integration of acoustic features and LDA model demonstrates a high accuracy in distinguishing depression among Chinese university students, suggesting its potential utility in rapid and large-scale depression screening. MFCC may serve as objective and valid features for the automated identification of depression on Chinese university campuses.

Objective:The aim of this study was to develop and validate a machine learning model to predict the risk of comorbid depression in asthma patients.Methods:We conducted a retrospective study of 2464 asthma patients with comorbid depression using National Health and Nutrition Examination Survey (NHANES) data. Feature selection was conducted using the Boruta algorithm and the Least Absolute Shrinkage and Selection Operator (LASSO). Eight machine learning algorithms, namely Decision Tree (DT), k-Nearest Neighbors (KNN), Light Gradient Booster Machine (LGBM), Logistic Regression (LR), Random Forest (RF), Support Vector Machine (SVM), eXtreme Gradient Boosting (XGBoost), and Multilayer Perceptron (MLP), were trained using 5-fold cross-validation methodology. Model performance was evaluated through various metrics such as area under the curve (AUC), accuracy, sensitivity, specificity, F1 score, and decision curve analysis (DCA). Interpretation was conducted using SHapley Additive exPlanations (SHAP) analysis, highlighting feature importance.Results:The training set comprised 1724 participants, while the validation set included 740 participants, with a depression prevalence of 14.45%. Significant predictors identified included hypertension, chronic obstructive pulmonary disease (COPD), stroke, sleep questionnaire (SLQ) scores, smoking status, Poverty Index Ratio (PIR), and educational level. The XGBoost model demonstrated superior performance compared with alternative machine learning (ML) algorithms, achieving an AUC of 0.750, an accuracy of 69.1%, a sensitivity of 68.2%, a specificity of 73.8%, and an F1 score of 79%. The SHAP method identified SLQ, PIR, and education level as the primary decision factors influencing the ML model’s predictions.Conclusion:The XGBoost model effectively predicts the risk of depression in asthma patients, serving as a valuable reference for early clinical identification and intervention.

Breast Cancer (BC) is counted as one of the most fatal diseases responsible for mortality in women. BC feels like a lump that can be felt different than other BC tissues. Cancer cells spread to other parts of the body by getting into the bloodstream or lymphatic system, which results in BC. It becomes crucial to develop and analyze methods to help understand BC and its trend affecting women. In this paper, we discussed different models to analyze BC tumors and classified them into benign and malignant. Benign tumors are referred to as ‘non-cancerous’ by many, and those that are ‘cancerous’ as malignant. The analysis was done on Wisconsin Breast Cancer Dataset (WBCD). We performed Logistic Regression (LR), Decision Tree(DT), Random Forest(RF), K-Nearest Neighbor (KNN), and Artificial Neural Network(ANN) on the dataset and compared their results. Our ANN model yielded the best result among many. All models were compared on various metrics which are Accuracy score, F1 score, Precision score, and Recall score.

Tree-based machine learning (ML) algorithms, such as Extra Trees (ET), Random Forest (RF), Gradient Boosting Machine (GBM), and XGBoost (XGB) are among the most widely used in early drug discovery, given their versatility and performance. However, models based on these algorithms often suffer from misclassification and reduced interpretability issues, which limit their applicability in practice. To address these challenges, several approaches have been proposed, including the use of SHapley Additive Explanations (SHAP). While SHAP values are commonly used to elucidate the importance of features driving models' predictions, they can also be employed in strategies to improve their prediction performance. Building on these premises, we propose a novel approach that integrates SHAP and features value analyses to reduce misclassification in model predictions. Specifically, we benchmarked classifiers based on ET, RF, GBM, and XGB algorithms using data sets of compounds with known antiproliferative activity against three prostate cancer (PC) cell lines (i.e., PC3, LNCaP, and DU-145). The best-performing models, based on RDKit and ECFP4 descriptors with GBM and XGB algorithms, achieved MCC values above 0.58 and F1-score above 0.8 across all data sets, demonstrating satisfactory accuracy and precision. Analyses of SHAP values revealed that many misclassified compounds possess feature values that fall within the range typically associated with the opposite class. Based on these findings, we developed a misclassification-detection framework using four filtering rules, which we termed "RAW", SHAP, "RAW OR SHAP", and "RAW AND SHAP". These filtering rules successfully identified several potentially misclassified predictions, with the "RAW OR SHAP" rule retrieving up to 21%, 23%, and 63% of misclassified compounds in the PC3, DU-145, and LNCaP test sets, respectively. The developed flagging rules enable the systematic exclusion of likely misclassified compounds, even across progressively higher prediction confidence levels, thus providing a valuable approach to improve classifier performance in virtual screening applications.

Background: Breast cancer is the most common malignancy among women worldwide, underscoring the importance of early detection and accurate prognostication. Machine learning (ML) has emerged as a promising approach, offering powerful tools for analyzing complex datasets in breast cancer prediction and diagnosis. Objective: This study evaluates the predictive performance of diverse ML algorithms for breast cancer classification using publicly available datasets, focusing on accuracy, interpretability, and generalizability. Methods: The dataset included clinical and demographic variables such as age, menopausal status, tumor size, and lymph node involvement. Data preprocessing addressed missing values and class imbalance, with the Synthetic Minority Oversampling Technique (SMOTE) applied to improve sensitivity for the minority class. Feature engineering involved interaction terms and scaling of numerical variables. Multiple ML models—Logistic Regression, Decision Tree, Random Forest, Gradient Boosting, Support Vector Machine (SVM), Naive Bayes, K-Nearest Neighbors (KNN), and Neural Networks—were trained and evaluated. Performance was measured using sensitivity, F1-score, and AUC-ROC. Model interpretability was enhanced with SHapley Additive exPlanations (SHAP). Results: Random Forest achieved the best performance with an AUC-ROC of 0.9751, followed by Gradient Boosting (0.9242) and Neural Networks (0.9254). Logistic Regression and SVM yielded comparable results (0.9005 and 0.9344). Ensemble models showed higher accuracy and generalizability, particularly on external validation. Tumor size and lymph node involvement emerged as key predictors. SMOTE improved sensitivity across models. Conclusion: This study demonstrates the potential of ML in breast cancer prediction, emphasizing the effectiveness of ensemble methods and interpretability tools. Future work should focus on integrating ML into clinical practice for earlier detection and personalized treatment.

IntroductionVaccination is critical for reducing childhood mortality, yet completion rates for the third dose of the pentavalent vaccine (Penta 3) in East Africa remain inadequate. This study aims to predict Penta 3 vaccination dropout using a stacking ensemble machine learning model with Demographic and Health Survey (DHS) data. The objective is to identify predictors of dropout and enhance intervention strategies.MethodsThe study utilized seven base machine learning algorithms to create a stacked ensemble model with three meta-learners: Random Forest (RF), Generalized Linear Model (GLM), and Extreme Gradient Boosting (XGBoost). The H2O package facilitated the development of base learners and the stacking of super learners. Feature selection (FS) and comparisons were performed using the LASSO and Boruta algorithms. The selected features were one-hot encoded, and ordinal encoding was applied where appropriate. Hyperparameter optimization (HPO) and comparisons were conducted using grid search and random search. Model performance was assessed using five key metrics, including accuracy and the area under the curve (AUC). SHAP (Shapley Additive Explanations) values were used to interpret the model outputs and identify influential predictors. The experimental design was employed to present the results.ResultsFour experiments were conducted to evaluate feature selection and HPO methods. All stacked ensemble models outperformed individual learners, with the XGBoost meta-learner optimized with grid search and LASSO FS achieving the highest performance: 93.9% accuracy and 99.4% AUC. While RF and GLM meta-learners were also evaluated, they were outperformed by the XGBoost meta-learner. SHAP analysis revealed key features influencing Penta 3 dropout, including the place of delivery, decision-making autonomy, the mother's level of earning, and healthcare access. Home delivery increased the risk of dropout, while postnatal care by midwives and health insurance coverage lowered dropout likelihood.Conclusion and recommendationThis study provides insights into the factors influencing Penta 3 vaccination dropout in East Africa. To reduce dropout rates, interventions should focus on enhancing maternal livelihood opportunities, improving healthcare access in rural areas, and promoting institutional deliveries.