Development of a pan-cancer prognostic neural network model
Principal Investigator
Name
Kyunghyun Paeng
Degrees
B.S.
Institution
Lunit
Position Title
Senior researcher
Email
About this CDAS Project
Study
PLCO
(Learn more about this study)
Project ID
PLCO-388
Initial CDAS Request Approval
Aug 10, 2018
Title
Development of a pan-cancer prognostic neural network model
Summary
As tumor micro-environment rises as a key to comprehend immune dynamics in various cancers, computational histopathology is regarded as a breakthrough approach to figure out what happens around the tumor before/during treatment. With this project, we would like to develop accurate computer algorithm powered by deep learning techniques to support physicians diagnose various cancers including lung, ovarian, and colorectal cancer and explore potential of imaging biomarkers based upon H&E stained images combined with patient profiles. We aim to train the model mainly with pathological lung, ovarian, and colorectal cancer images, and images from other cancer sites can be used as well, if available, in order to develop more robust network model working across all types of tumors.
Aims
1) Detection of tissue compartments using a convolutional neural network across various types of cancers - mainly lung, ovarian, and colorectal cancer
2) Development of a prognostic machine learning model based upon H&E stained slides
3) Exploring the potential of tumor type specific imaging biomarkers
Collaborators
All members at the R&D center of Lunit Inc.