Skip to Main Content

An official website of the United States government

Principal Investigator
Name
Xuefei Li
Degrees
Ph.D
Institution
CAS Key Laboratory for Quantitative Engineering Biology, Shenzhen Institute of Synthetic Biology, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences
Position Title
Associate Professor
Email
About this CDAS Project
Study
NLST (Learn more about this study)
Project ID
NLST-807
Initial CDAS Request Approval
Jul 2, 2021
Title
Investigating the association between cancer-cell epithelial-to-mesenchymal transition and the location of immune cells in lung adenocarcinoma based on pathology images
Summary
It has been demonstrated that for various types of cancer, both the epithelial-to-mesenchymal transition (EMT) of cancer cells and the infiltration of cytotoxic T lymphocytes can independently predict the prognosis of patients. However, the association between the EMT-status of cancer cells and the infiltration of lymphocytes in solid tumors is still in debate. Quantifying the correlation between the two and elucidating the underlying mechanisms could be helpful for the rational design of combination therapies. In this study, for lung adenocarcinoma, we plan to develop an automated pathology image processing pipeline, which includes the segmentation of tumor-cells/stromal-cells/lymphocytes based on convolutional neural network, the evaluation of cancer-cell EMT status based on the tumor-cell morphology, and the statistical analysis of the association between cancer-cell EMT status and the spatial distribution of lymphocytes. In addition, we will perform quantitative analyses to elucidate the spatial correlations between the cancer-cell EMT and the locations of lymphocytes, and more importantly investigate the effects of such correlation on the prognosis of patients.
Aims

1. Develop an automated image analysis pipeline for lung adenocarcinoma pathology images, and identify individual tumor cells, stromal cells, and lymphocytes using deep-learning based methods.
2. Identify the EMT status of cancer cells based on the shape of their nucleus using deep-learning based methods.
3. Calculate the infiltration pattern of lymphocytes into tumor islets.
4. Evaluate the spatial correlation between cancer-cell EMT and the infiltration patterns of lymphocytes.
5. Develop a deep-learning based method to predict patient prognosis using processed pathological images with spatial information of tumor cells and lymphocytes as well as the EMT status of cancer cells.

Collaborators

Herbert Levine, Northeastern University
Mohit Kumar Jolly, Indian Institute of Science
Ju Cui, Beijing Hospital