Jul 2026· Network Modeling Analysis in Health Informatics and Bioinformatics· Vol 15· 0 citations· 34 references
Computer Science
TL;DR
A multimodal Markov Blanket (MB)-based feature selection framework that performs local structure learning across seven data modalities and introduces a composite dependency-ranking fallback when sparse neighborhoods lead to data starvation is developed.
A multi-tier, explainable AI framework designed to risk-stratify patients and predict overall survival using clinical and genomic covariates is developed and demonstrates that explainable machine learning models can robustly predict survivability and highlight actionable features for oncology dashboards.
Abstract Motivation Accurate survival prediction is crucial for personalized cancer treatment but remains challenging for rare cancers due to limited data. Most deep learning models require large training datasets, which are unavailable for rare cancer types, creating a significant clini-cal bottleneck. Results We propose MoESurv, a zero-sample survival prediction framework that leverages a mix-ture-of-experts architecture to extract generalizable prognostic patterns from pan-cancer data. The model integrates shared experts, cancer-specific experts, and routing experts within an autoencoder to disentangle common and type-specific survival features. Evaluated on seven rare TCGA cancer types, MoESurv achieved state-of-the-art performance, improving the average C-index by 4 percentage points over the best baseline. Further-more, external validation across diverse populations and independent cohorts—including a Chinese glioma cohort (CGGA mRNAseq_693, C-index = 0.7433), a rare GBM IDH-mutant subtype (C-index = 0.8064), and the pan-cancer PCAWG cohort (C-index = 0.7090)—demonstrated that MoESurv possesses the most robust predictive per-formance, highlighting its generalizability. MoESurv also effectively stratified high- and low-risk patient groups and identified potential survival-associated genes, demonstrating both clinical utility and biological interpretability. Availability The code is freely available at https://github.com/HuaYC666/MoESurv and https://zenodo.org/records/20785500.
This work proposes MultiSigBERT, a unified framework for multimodal sequential survival modeling in oncology based on path signature representations that achieves a concordance index of 0.743 on an independent test set, demonstrating the benefit of jointly modeling multimodal temporal dynamics together with patient-level geometric structure for survival prediction.
Paul Minchella, Stéphane Chrétien, Guillaume Metzler et al.· 0 citations
Breast cancer remains a major global health challenge, requiring accurate and interpretable diagnostic systems for reliable clinical deployment. This paper proposes an Explainable Hybrid Machine Learning (XML) framework that integrates advanced feature extraction with interpretable classification techniques for early breast cancer detection and staging. Using benchmark datasets including WDBC, BreakHis, and CBIS-DDSM, the framework applies CLAHE-based preprocessing, PCA, and SHAP-driven Recursive Feature Elimination (SHAP-RFE) to generate an optimized Hybrid Feature Vector (HFV). Experimental results demonstrate a classification accuracy of 96.84% and sensitivity of 97.12%, outperforming conventional machine learning models while reducing overfitting. The framework further supports multi-class staging aligned with AJCC TNM criteria using a dual Explainable AI subsystem combining Grad-CAM++ and SHAP for visual and mathematical interpretability. Inclusion of molecular pathways such as MAPK and PI3K-Akt improved predictive reliability by 9.2%. The proposed system offers a robust, transparent, and clinically auditable solution for personalized breast cancer diagnosis and treatment planning.
Unknown authors· International journal of com...· 0 citations
Accurate breast cancer prognosis remains a major challenge in precision oncology due to tumor heterogeneity and the complexity of integrating high-dimensional multi-omics data. Although multimodal learning approaches have improved predictive performance by combining clinical and molecular information, many existing methods rely on a single ensemble strategy that remains susceptible to prediction variance and limited robustness in high-dimensional, low-sample-size biomedical datasets. This study investigated whether integrating complementary ensemble strategies within a unified multimodal framework could improve the robustness and predictive performance of breast cancer prognosis. A heterogeneous multimodal ensemble framework was developed in which stacking was used to integrate complementary information from clinical, gene expression, and copy number variation (CNV) data through meta-learning, while bagging was incorporated to stabilize the meta-learning process via bootstrap aggregation. The outputs of the stacking and bagging branches were combined using weighted probability fusion. The framework was evaluated on the METABRIC breast cancer cohort and compared with unimodal models and a conventional stacking ensemble using an independent test set and stratified tenfold cross-validation. The proposed hybrid framework achieved a ROC-AUC of 0.936, outperforming unimodal clinical and molecular models (ROC-AUC = 0.8140.885) and the conventional stacking ensemble (ROC-AUC = 0.898). Stratified tenfold cross-validation further demonstrated consistent improvements in mean ROC-AUC, recall, F1-score, balanced accuracy, and Matthews correlation coefficient, indicating improved robustness and stable performance across the internal validation folds. On the independent test set, the hybrid framework reduced false-negative predictions and increased sensitivity relative to the stacking ensemble, demonstrating a more favorable balance between identifying high-risk patients and maintaining overall predictive performance. Rather than introducing a new ensemble algorithm, this study demonstrates that assigning complementary roles to stacking multimodal information integration and bagging for prediction stabilization provides an effective and robust framework for multi-omics breast cancer prognosis. The proposed hybrid strategy consistently improved predictive performance and robustness compared with conventional stacking while demonstrating stable performance across internal validation, supporting the use of complementary ensemble paradigms for multimodal prediction in precision oncology.
Reza Bozorgpour, Mohammadreza Soltany Sadrabadi· Clinical Cancer Bulletin· 0 citations
Lung cancer remains one of the leading causes of cancer-related mortality worldwide, primarily due to challenges in early-stage detection and accurate risk stratification. Conventional machine learning models often exhibit limited interpretability and reduced predictive capability when modelling complex, non-linear relationships among demographic and lifestyle risk factors. Therefore, robust, explainable, and methodologically rigorous predictive frameworks are required to support reliable clinical decision-making. This study proposes a systematically validated explainable ensemble learning framework for lung cancer stage prediction by integrating optimized stacking and voting strategies. A leakage-free machine learning pipeline comprising data preprocessing, normalization, recursive feature elimination (RFE), and GridSearchCV-based hyperparameter optimization was developed. Multiple machine learning classifiers were combined through optimized ensemble learning, while SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME) were employed to provide complementary global and local interpretability. Model performance was evaluated using an independent hold-out test set, stratified 10-fold cross-validation, 95% confidence intervals, non-parametric statistical analysis using the Friedman aligned-ranks test with Holm post-hoc analysis, and a component-wise ablation study. The optimized stacking model (STB) achieved the best overall predictive performance, with an accuracy of 99.52%, precision of 99.57%, recall of 99.50%, F1-score of 99.50%, and AUC of 100%. Statistical validation confirmed the robustness and comparative performance of the proposed framework, while the ablation study verified the contribution of feature selection, ensemble learning, and hyperparameter optimization. SHAP and LIME consistently identified passive smoking (PS), obesity (OB), coughing of blood (CB), wheezing (WH), and fatigue (FT) as the most influential predictors of lung cancer stage. The proposed explainable ensemble learning framework provides accurate, robust, and interpretable lung cancer stage prediction through the integration of leakage-free model development, optimized ensemble learning, explainable artificial intelligence, statistical validation, and component-wise ablation analysis. These findings demonstrate its potential as a reliable decision-support framework while highlighting the need for future validation using independent multi-center clinical datasets.
M. Alfuraydan, Shahid Mohammad Ganie, Ehab Seedahmed et al.· Discover Computing· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
