Survival Analysis with Whole Slide Images based on Deep Learning Technique
Principal Investigator
Name
Lei Fan
Degrees
M.D.
Institution
UNSW Sydney
Position Title
Ph.D candidate
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-734
Initial CDAS Request Approval
Dec 11, 2020
Title
Survival Analysis with Whole Slide Images based on Deep Learning Technique
Summary
Advances in WSIs technologies provide physicians with vast amounts of data, yet prognostication remains largely subjective, leading to suboptimal clinical management. Because manual assessments with histopathological images are high time-consuming and subjective, especially by pathologists who have varying levels of experience.
In this project, a whole pipeline will be developed to make survival analysis with WSIs based on deep learning techniques. Specifically, a comprehensive method based on weakly-supervision technique will be proposed to select RoI region from WSIs. And then an end-to-end architecture will be built to implement automatic features extraction, where region will put into deep learning model to generate survival prediction.
In this project, a whole pipeline will be developed to make survival analysis with WSIs based on deep learning techniques. Specifically, a comprehensive method based on weakly-supervision technique will be proposed to select RoI region from WSIs. And then an end-to-end architecture will be built to implement automatic features extraction, where region will put into deep learning model to generate survival prediction.
Aims
1. construct a framework that can analyze whole slide images automatically
2. make survival analysis with whole slide images based on deep learning, especially weakly supervison technique
3. build a unified system that can implement survival prediction from various types of cancers
Collaborators
Song Yang, UNSW Sydney
Related Publications
-
Cancer Survival Prediction From Whole Slide Images With Self-Supervised Learning and Slide Consistency.
Fan L, Sowmya A, Meijering E, Song Y
IEEE Trans Med Imaging. 2023 May; Volume 42 (Issue 5): Pages 1401-1412 PUBMED