Representation Learning of Histopathology Images using Graph Neural Networks
This work addresses the need for higher precision in diagnostic pathology for lung cancer patients, but it is incremental as it builds on existing methods with specific improvements.
The authors tackled the problem of classifying lung cancer subtypes from whole slide images by proposing a two-stage framework that uses graph neural networks to aggregate patch information, achieving state-of-the-art accuracy of 88.8% and AUC of 0.89.
Representation learning for Whole Slide Images (WSIs) is pivotal in developing image-based systems to achieve higher precision in diagnostic pathology. We propose a two-stage framework for WSI representation learning. We sample relevant patches using a color-based method and use graph neural networks to learn relations among sampled patches to aggregate the image information into a single vector representation. We introduce attention via graph pooling to automatically infer patches with higher relevance. We demonstrate the performance of our approach for discriminating two sub-types of lung cancers, Lung Adenocarcinoma (LUAD) & Lung Squamous Cell Carcinoma (LUSC). We collected 1,026 lung cancer WSIs with the 40$\times$ magnification from The Cancer Genome Atlas (TCGA) dataset, the largest public repository of histopathology images and achieved state-of-the-art accuracy of 88.8% and AUC of 0.89 on lung cancer sub-type classification by extracting features from a pre-trained DenseNet