Lung and colon cancers remain major contributors to cancer-related morbidity and mortality worldwide. Accurate histopathological image classification can support early diagnosis and improve clinical decision-making. In this study, we propose a hybrid deep learning framework for histopathological image classification that combines dual transfer learning with Neighbour Feature Attention-based Pooling (NFP). The framework employs DenseNet-169 and InceptionResNet101-V2 as complementary pretrained backbones, and the extracted high-level features are fused through a feature-level concatenation strategy for end-to-end classification. NFP is integrated into both backbones to preserve spatial dependencies and contextual tissue information that are often lost in standard pooling operations. The proposed model is evaluated on the LC25000 dataset, which contains five classes: Lung Adenocarcinoma, Lung Squamous Cell Carcinoma, Lung Benign, Colon Adenocarcinoma, and Colon Benign. Experimental results show that the proposed framework achieves an accuracy of 96.3%, with precision, recall, F1-score, and AUC values of 0.96, 0.97, 0.96, and 0.98, respectively. These results demonstrate that the proposed method is effective and robust for automated lung and colon cancer histopathological image classification and has potential as a diagnostic support tool.
Das et al. (Fri,) studied this question.