Weakly-supervised learning to develop an AI-based risk-stratification of PLCO cancers

Principal Investigator

Name
Okyaz Eminaga

Degrees
M.D., Ph.D.

Institution
Okyaz Eminaga

Position Title
Instructor

Email
okyaz.eminaga@gmail.com

About this CDAS Project

Study
PLCO (Learn more about this study)

Project ID
PLCOI-1111

Initial CDAS Request Approval
Nov 28, 2022

Title
Weakly-supervised learning to develop an AI-based risk-stratification of PLCO cancers

Summary
The current study aims to examine the capability of AI to learn features from PLCO cancers and develop risk-stratification tools that are prognostic. We will further evaluate AI on independent dataset to validate our concept.

Aims

1) To build a framework to process and tiles histology images
2) To train novel AI models to learn features association with cancer-specific survival
3) To validate AI models on independent datasets with prognostic and molecular biology informations.

Collaborators

Mahmoud Abbas, University of Muenster.