Lung cancer multi-year risk prediction
Principal Investigator
Name
Zhiying Liang
Degrees
Master
Institution
ShanghaiTech University
Position Title
Student
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-1279
Initial CDAS Request Approval
Jul 15, 2024
Title
Lung cancer multi-year risk prediction
Summary
This project focuses on implementing a risk prediction model for predicting multi-year lung cancer using a single low-dose CT image of the lung, specifically using artificial intelligence algorithms. The CT image is fed into a 3D visual coder to get the CT image features of lung cancer, and then a classifier is used to predict whether the CT image is cancerous or non-cancerous, and then such a prediction is made for each year, which in turn leads to a multi-year risk prediction.There are multi-year follow-up results in the NLST dataset, and there is a record of multi-year follow-up of lung cancer patients, which provides good data support for this project to realise this artificial intelligence model.
Aims
The goal of this project is mainly to get better prediction accuracy for lung cancer and non-cancer and to get multi-year cancer risk prediction results using assessment metrics such as C-index, AUC, sensitivity, and specificity, etc., and further analyses using statistics, such as calculating confidence intervals. The lung region in the CT image of interest to the model was visualised using the grad-cam algorithm.
Collaborators
Zhiying Liang (ShanghaiTech University )