Towards Explainable Graph Representations in Digital Pathology

Explainability of machine learning (ML) techniques in digital pathology (DP) is of great significance to facilitate their wide adoption in clinics. Recently, graph techniques encoding relevant biological entities have been employed to represent and assess DP images. Such paradigm shift from pixel-wise to entity-wise analysis provides more control over concept representation. In this paper, we introduce a post-hoc explainer to derive compact per-instance explanations emphasizing diagnostically important entities in the graph. Although we focus our analyses to cells and cellular interactions in breast cancer subtyping, the proposed explainer is generic enough to be extended to other topological representations in DP. Qualitative and quantitative analyses demonstrate the efficacy of the explainer in generating comprehensive and compact explanations.

[1]  Max Welling,et al.  Semi-Supervised Classification with Graph Convolutional Networks , 2016, ICLR.

[2]  Fabio A. González,et al.  A Deep Learning Architecture for Image Representation, Visual Interpretability and Automated Basal-Cell Carcinoma Cancer Detection , 2013, MICCAI.

[3]  Geoffrey E. Hinton,et al.  Distilling the Knowledge in a Neural Network , 2015, ArXiv.

[4]  Zhipeng Jia,et al.  Large scale tissue histopathology image classification, segmentation, and visualization via deep convolutional activation features , 2017, BMC Bioinformatics.

[5]  Jure Leskovec,et al.  How Powerful are Graph Neural Networks? , 2018, ICLR.

[6]  Samuel S. Schoenholz,et al.  Neural Message Passing for Quantum Chemistry , 2017, ICML.

[7]  Ming Y. Lu,et al.  Weakly Supervised Prostate Tma Classification Via Graph Convolutional Networks , 2019, 2020 IEEE 17th International Symposium on Biomedical Imaging (ISBI).

[8]  Xavier Bresson,et al.  Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering , 2016, NIPS.

[9]  D. Brat,et al.  Predicting cancer outcomes from histology and genomics using convolutional networks , 2017, Proceedings of the National Academy of Sciences.

[10]  Olaf Hellwich,et al.  Cell nuclei attributed relational graphs for efficient representation and classification of gastric cancer in digital histopathology , 2016, SPIE Medical Imaging.

[11]  Saeed Hassanpour,et al.  Looking Under the Hood: Deep Neural Network Visualization to Interpret Whole-Slide Image Analysis Outcomes for Colorectal Polyps , 2017, 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).

[12]  Klaus-Robert Müller,et al.  Resolving challenges in deep learning-based analyses of histopathological images using explanation methods , 2019, Scientific Reports.

[13]  Natalia Gimelshein,et al.  PyTorch: An Imperative Style, High-Performance Deep Learning Library , 2019, NeurIPS.

[14]  Hao Chen,et al.  A Multi-Organ Nucleus Segmentation Challenge , 2020, IEEE Transactions on Medical Imaging.

[15]  Jure Leskovec,et al.  GNNExplainer: Generating Explanations for Graph Neural Networks , 2019, NeurIPS.

[16]  Andreas Holzinger,et al.  Towards the Augmented Pathologist: Challenges of Explainable-AI in Digital Pathology , 2017, ArXiv.

[17]  Alexander Binder,et al.  On Pixel-Wise Explanations for Non-Linear Classifier Decisions by Layer-Wise Relevance Propagation , 2015, PloS one.

[18]  Jin Tae Kwak,et al.  Hover-Net: Simultaneous segmentation and classification of nuclei in multi-tissue histology images , 2018, Medical Image Anal..

[19]  Cigdem Demir,et al.  The cell graphs of cancer , 2004, ISMB/ECCB.

[20]  Henning Müller,et al.  Regression Concept Vectors for Bidirectional Explanations in Histopathology , 2018, MLCN/DLF/iMIMIC@MICCAI.

[21]  Pheng-Ann Heng,et al.  CGC-Net: Cell Graph Convolutional Network for Grading of Colorectal Cancer Histology Images , 2019, 2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW).

[22]  Pietro Liò,et al.  Graph Attention Networks , 2017, ICLR.

[23]  Shrey Gadiya,et al.  Histographs: Graphs in Histopathology , 2019, Medical Imaging: Digital Pathology.