Lymphatic Node Cancer Classification using Advanced Hybrid Deep Neural Architecture

Metastatic cancer, characterized by the spread of cancer cells from the primary site to other parts of the body, poses significant diagnostic challenges. Accurate detection in histopathological images is crucial for effective patient management. In this study, we present a novel hybrid model combining ResNet50, self-attention mechanisms, and Gated Recurrent Units (GRUs) to enhance binary classification accuracy for metastatic cancer detection. Benchmarking against CNN-GRU, CNN-LSTM, and AlexNet-GRU models, our approach demonstrated superior performance on two datasets. For the Histopathologic Cancer Detection PCam Dataset, our model achieved 99.7% accuracy, 99.57% precision, 99.2% sensitivity, and 99.57% specificity. For the BreakHis Dataset, it attained 99.2% accuracy, 98.69% precision, 99.52% sensitivity, and 97.22% specificity. These results highlight our model’s potential to significantly reduce diagnostic errors and support pathologists in making more accurate diagnoses, outperforming existing models in the field.