Abstract:A clinically comparable Convolutional Neural Network framework-based technique for performing automated classification of cancer grades and tissue structures in hematoxylin and eosin-stained colon histopathological images is proposed in this paper. It comprised of Enhanced Convolutional Learning Modules (ECLMs), multi-level Attention Learning Module (ALM), and Transitional Modules (TMs). The ECLMs perform a dual mechanism to extract multi-level discriminative spatial features and model cross-channel correlations with fewer computations and effectual avoidance of vanishing gradient issues. The ALM performs focus-refinement through the channel-wise elemental attention learning to accentuate the discriminative channels of the features maps specifically belonging to the important pathological regions and the scale-wise attention learning to facilitate recalibration of features maps at diverse scales. The TMs concatenate the output of these two modules, infuse deep multi-scalar features and eliminate resolution degradation issues. Varied pre-processing techniques are further employed to improvise the generalizability of the proposed network. For performance evaluation, four diverse publicly available datasets (Gland Segmentation challenge(GlaS), Lung Colon(LC)-25000, Kather_Colorectal_Cancer_Texture_Images (Kather-5k), and NCT_HE_CRC_100K(NCT-100k)) and a private dataset Hospital Colon(HosC) are used that further aids in building network invariance against digital variability that exists in real-clinical data. Also, multiple pathologists are involved at every stage of the proposed research and their verification and approval are taken for each step outcome. For the cancer grade classification, the proposed model achieves competitive results for GlaS (Accuracy(97.5%), Precision(97.67%), F1-Score(97.67%), and Recall(97.67%)), LC-25000 (Accuracy(100%), Precision(100%), F1-Score(100%), and Recall(100%)), and HosC(Accuracy(99.45%), Precision(100%), F1-Score(99.65%), and Recall(99.31%)), and while for the tissue structure classification, it achieves results for Kather-5k(Accuracy(98.83%), Precision(98.86%), F1-Score(98.85%), and Recall(98.85%)) and NCT-100k(Accuracy(97.7%), Precision(97.69%), F1-Score(97.71%), and Recall(97.73%)). Furthermore, the reported activation mappings of Gradient-Weighted Class Activation Mappings(Grad-CAM), Occlusion Sensitivity, and Local Interpretable Model-Agnostic Explanations (LIME) evidence that the proposed model can itself learn the similar patterns considered pertinent by the pathologists exclusive of any prerequisite for annotations. In addition, these visualization results are inspected by multiple expert pathologists and provided with a validation score as (GlaS(9.251), LC-25000(9.045), Kather-5k(9.248), NCT-100k(9.262), and HosC(9.853)). This model will provide a secondary referential diagnosis for the pathologists to ease their load and aid them in devising an accurate diagnosis and treatment plan.

HACT-Net: A Hierarchical Cell-to-Tissue Graph Neural Network for Histopathological Image Classification

Hierarchical graph representations in digital pathology

A Scalable Graph-Based Framework for Multi-Organ Histology Image Classification

A Multi-scale Graph Network with Multi-head Attention for Histopathology Image Diagnosisn

A convolution neural network with multi-level convolutional and attention learning for classification of cancer grades and tissue structures in colon histopathological images

CCF-GNN: A Unified Model Aggregating Appearance, Microenvironment, and Topology for Pathology Image Classification

Histographs: Graphs in Histopathology

Single-Cell Spatial Analysis of Histopathology Images for Survival Prediction Via Graph Attention Network

Histopathological Image Classification with Cell Morphology Aware Deep Neural Networks

PathoGraph: An Attention-Based Graph Neural Network Capable of Prognostication Based on CD276 Labelling of Malignant Glioma Cells

Cells are Actors: Social Network Analysis with Classical ML for SOTA Histology Image Classification

Heterogeneous graphs model spatial relationships between biological entities for breast cancer diagnosis

A Convolutional Neural Network and Graph Convolutional Network Based Framework for Classification of Breast Histopathological Images

Spatial-Hierarchical Graph Neural Network with Dynamic Structure Learning for Histological Image Classification

Graph Convolutional Neural Networks for Histological Classification of Pancreatic Cancer

A Hierarchical Graph V-Net with Semi-Supervised Pre-Training for Histological Image Based Breast Cancer Classification.

A Survey on Graph-Based Deep Learning for Computational Histopathology

Hierarchical Deep Convolutional Neural Networks for Multi-category Diagnosis of Gastrointestinal Disorders on Histopathological Images

DeepHistoNet: A robust deep-learning model for the classification of hepatocellular, lung, and colon carcinoma

A GNN-based Network for Tissue Semantic Segmentation in Histopathology Image

A Hierarchical Hypergraph Attention Network for Survival Analysis from Pathological Images