Breast Cancer: Aetiology, Diagnosis, Prevention, Treatment, and the Transformative Role of Artificial Intelligence

Breast cancer is a life-threatening disease affecting women, and it needs effective and efficient diagnostic procedures to detect it early enough to ensure effective management. Artificial Intelligence (AI) offers significant promise for improving breast cancer detection and treatment; however, its application remains constrained by methodological, ethical, and infrastructural limitations. This paper examines a hybrid anomaly-detection model applied to the breast cancer dataset from Kaggle. The dataset is preprocessed and class-balanced using Mean Absolute Error (MAE) thresholding and SMOTE techniques. Twelve ML and DL models were trained and compared using the dataset in terms of typical evaluation metrics. The highest diagnostic performance was obtained with the LightGBM classifier with 0.9912 Accuracy, 0.9880 F1 Score, and a Specificity (1.0000). The Results indicate the usefulness of ensemble techniques in high-accuracy biomedical anomaly detection. Persistent issues with insufficient dataset diversity, model interpretability, and clinical standards should be resolved to facilitate the deployment of reliable AI systems in the fields of breast cancer detection, prevention, and therapy.