Background and objective <p>Breast cancer is a leading cause of cancer-related death among women globally. Early and accurate detection is crucial for improving survival rates. However, traditional mammography often suffers from diagnostic variability because of its dependence on radiologists’ interpretations. This study aimed to enhance breast cancer classification accuracy by integrating radiomic features with deep learning-derived features, particularly addressing the challenges posed by limited datasets.</p> Methods <p>A hybrid classification framework was developed by combining radiomic and deep-learning features. Recursive Feature Elimination with cross-validation is used to select the most relevant features. An ensemble of machine learning classifiers was employed for classification, and the performance was evaluated using accuracy, precision, sensitivity, specificity, and ROC-AUC score. The model was tested on a mammogram image dataset with limited sample size.</p> Results <p>The ensemble classifier achieved 98.43% accuracy, 98.50% precision, 96.77% sensitivity and specificity, and an ROC-AUC score of 0.99. These metrics demonstrate substantial improvements over traditional approaches, with enhanced diagnostic accuracy and reduced variability.</p> Conclusions <p>The proposed integration of radiomics and deep learning features provides a reliable and accurate approach for breast cancer detection, even with limited data. This method offers a valuable decision-support tool for radiologists and has the potential to improve diagnostic consistency in clinical settings, especially in regions with limited access to expert radiological assessments.</p>