Data-driven Optimal Personalized Early Screening Strategies for Lung Cancer
Principal Investigator
Name
Jianjin Yue
Degrees
Ph.D student
Institution
Sichuan University
Position Title
Ph.D student
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-1374
Initial CDAS Request Approval
Jan 6, 2025
Title
Data-driven Optimal Personalized Early Screening Strategies for Lung Cancer
Summary
This project aims to establish a comprehensive end-to-end framework for lung nodule and lung cancer screening. We first develop a deep learning model suitable for multimodality to extract pathological features of lung nodules and lung cancer from CT images. Secondly, we build a deep learning model capable of handling temporal information to characterize the evolution trajectory of pathological features and predict future development. Finally, we solve for the optimal personalized early screening strategy for specific patients by establishing a POMDP model. The NLST dataset will be used for external validation to ensure the accuracy, robustness, and generalizability of the proposed deep learning models. This research can provide the optimal early screening strategy for lung cancer patients, help improve patient outcomes, and contribute to enhancing medical service efficiency and reducing medical costs.
Aims
1. Extract patient pathological feature information from CT images and demographic data using deep learning methods.
2. Characterize the evolution trajectory of patient pathological features using deep learning methods.
3. Determine the optimal screening strategy for patients using operations research methods.
Collaborators
Jianjin Yue, Business school, Sichuan University
Mengzhuo Guo, Business school, Sichuan University
Li Luo, Business school, Sichuan University