Medical Image Based Prognosis for Lung Cancer Patients
Principal Investigator
Name
Hongda Zhang
Degrees
Ph.D.
Institution
Nanjing University
Position Title
Postdoctoral fellow
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-994
Initial CDAS Request Approval
Dec 8, 2022
Title
Medical Image Based Prognosis for Lung Cancer Patients
Summary
An accurate prediction of survival will help guide the treatment planning for cancer patients. Various cancer staging play a big role in prognosis. The process is done manually by experienced medical doctors. However, it is time consuming and the results are usually objective. Thus, unified models which can not only extract prognosis related features from medical images, but also make accurate prediction of the survival of the cancer patients, are in demand. In recent years, deep learning methods are developed for medical image analysis and prognosis. In order to train such models, a large amount of multimodal medical images and clinical data are required. This study will utilize both pathology and radiology images and clinical data to develop a combined deep learning model and Bayesian survival model for accurately predicting the survivor function of lung cancer patients.
Aims
-develop a combined deep learning model and Bayesian survival model for prognosis of lung cancer patients.
-identify prognosis relevant features in both pathology and radiology images.
-select clinical factors relevant to prognosis.
-validate the performance of the model with missing values.
Collaborators
Xiaoping Yang, PhD, Nanjing University