Making Sense of Breast Cancer Grades: A New Way Using Multiple Models

Breast cancer grades, which are determined by cell structure, are crucial for planning treatment. Many researchers rely on deep learning (DL) models for this task. However, DL models can be a bit of a mystery. It's unclear which features they use to make accurate predictions. A recent study introduced a new model called Grade Differentiation Integrated Model (GradeDiff-IM). This model uses several machine learning (ML) models to classify breast cancer grades (G1, G2, and G3). It also considers important biological features and ranks them to help make decisions. After that, it uses deep learning models to analyze histopathological images and compares the results with the ML models. Instead of using just one model, GradeDiff-IM combines multiple models using a technique called stacking. This approach helps improve the accuracy of grade classification. The highest accuracy was achieved with grades G1 (98. 2%), G2 (97. 6%), and G3 (97. 5%). The study found that the ML ensemble model was more accurate than the DL models. This shows that using multiple models together can lead to better results than using just one.