Predicting Early- and Late-Stage Breast Cancer via Clinical–Genomic using Feature Fusion and Explainable Ensemble Learning
International Journal of Drug Delivery Technology · 01.06.2026 · 16 (42s)
DOI: 10.25258/ijddt.16.42s.85 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Drug Delivery Technology, online erschienen am 01.06.2026. Verfasst von Sonali Mondal Das und 1 weiteren Autor:innen.
Abstract (Originalsprache)
Correctly identifying what stage breast cancer is at is extremely valuable to doctors since it will improve their ability to make the appropriate diagnosis or treatment. To achieve this, we created a machine-learning system that utilized a combination of clinical information from the METABRIC data set along with a vast variety of Gene Expression features to classify the stage of breast cancers into an early (I & II) or late (III & IV) classification system. We were very diligent in our efforts to ensure that we were not overlooking an area of importance; we encoded, normalized, and achieved balance through the use of Tomek-SMOTE Balancing and stratified splitting. We tested various models - XGBoost, Random Forest, SVM, and Voting and Stacking Classifiers - and measured their performance by ROC-AUC, PR-AUC, and F1-Score to provide the most accurate representation of performance. For the early-stage breast cancer samples, there were marked differences in performance, and XGBoost achieved an area under the ROC curve of 0.9967 and an area under the precision/recall curve of 0.9981 which is indicative of a highly successful model. In late-stage breast cancer samples, SVM performed the best; however, its area under the ROC curve only reached 0.5772 which signifies that performance was not quite as successful. We utilized SHAP to gain insight into clinical and genomic markers, and it became evident that only a small number of markers have a major influence on breast cancer classification. Overall, this study indicates that for improved results on the classification of breast cancer, it is important to approach & develop models for each of the four stages separately while utilizing a combination of Clinical and Genomic data.
Zitieren
APA
Das, S. M., & Mondal, A. C. (2026). Predicting Early- and Late-Stage Breast Cancer via Clinical–Genomic using Feature Fusion and Explainable Ensemble Learning. International Journal of Drug Delivery Technology, 16(42s). https://doi.org/10.25258/ijddt.16.42s.85
Vancouver
Das SM, Mondal AC. Predicting Early- and Late-Stage Breast Cancer via Clinical–Genomic using Feature Fusion and Explainable Ensemble Learning. International Journal of Drug Delivery Technology. 2026;16(42s). doi:10.25258/ijddt.16.42s.85
BibTeX
@article{Das2026,
author = {Das, Sonali Mondal and Mondal, Abhoy Chand},
title = {Predicting Early- and Late-Stage Breast Cancer via Clinical–Genomic using Feature Fusion and Explainable Ensemble Learning},
journal = {International Journal of Drug Delivery Technology},
year = {2026},
volume = {16},
number = {42s},
doi = {10.25258/ijddt.16.42s.85}
}
EUVIMED