A Graph-transformer for Lung Subtype Classification on Whole Slide Images
Principal Investigator
Name
Yuxin Kang
Degrees
Ph.D.
Institution
Northwest University
Position Title
Docotor
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-1021
Initial CDAS Request Approval
Mar 8, 2023
Title
A Graph-transformer for Lung Subtype Classification on Whole Slide Images
Summary
Computational pathology, which entails the analysis of digitized pathology slides, is gaining increased attention over the past few years. The sheer size of a single whole slide image (WSI) typically can exceed a gigabyte, so traditional image analysis routines may not be able to fully process all this data in an efficient fashion. Most of these methods rely on systematic breakdown of WSIs into image patches, followed by development of deep neural networks at patch-level and integration of outcomes on these patches to create overall WSI-level estimates. However, patch-based methods introduce label noise during training by assuming that each patch is independent with the same label as the WSI and neglect overall WSI-level information that is significant in disease grading. Here we aim to propose a graph-transformer classifier that fuses a graph-based representation of an WSI and a vision transformer for processing pathology images, to predict lung subtype classification on WSIs.
Aims
We aim to develope a graph-based vision transformer for digital pathology called GTP that leverages a graph representation of pathology images and the computational efficiency of transformer architectures to perform WSI-level analysis. To build the GTP framework, we constructed a graph convolutional network by embedding image patches in feature vectors using contrastive learning, followed by the application of a vision transformer to predict a WSI-level label. We want to collect 1000 WSIs and clinical data to develope a model that could distinguish between normal, LUAD, and LSCC WSIs.
Collaborators
Xuan Zhao, Northwest University
Mo Zhou, Northwest University