Breast cancer represents the leading malignancy across the globe and is characterized by a continuous rise in both incidence and mortality rates each year. This study introduces an innovative predictive framework for determining appropriate breast cancer interventions—surgery, radiotherapy, and chemotherapy—using a multi-output classifier that leverages algorithms such as CatBoost, XGBoost, NN, and NN Binary. To address the urgent requirement for reliable therapeutic guidance, the developed model aims to substantially improve the accuracy of treatment outcome predictions. It performs exceptionally well in estimating surgical interventions, particularly with Neural Networks (both NN and NN Binary), delivering 97% accuracy and 98% F1-scores. Radiotherapy predictions yield a moderate overall accuracy of around 63%, with strong recall reaching up to 84%. For chemotherapy, the model maintains stable accuracy and precision levels of 82%, with AUC-ROC scores as high as 89%, reflecting superior classification capability. By merging multi-output classification techniques with state-of-the-art algorithms, this work aims to generate more patient-specific treatment forecasting tools, supporting the shift toward personalized oncology strategies and meeting the growing expectations of precision medicine in cancer management.
