Extraction and screening of early-onset non-small cell carcinoma patients with CT influence to predict tumor consistency and mutation
Principal Investigator
Name
明威 陈
Degrees
bachelor
Institution
Guilin University of Electronic Technology
Position Title
Student
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-803
Initial CDAS Request Approval
Jun 14, 2021
Title
Extraction and screening of early-onset non-small cell carcinoma patients with CT influence to predict tumor consistency and mutation
Summary
Due to the limitation of the range of the receptive field, the 3D U-Net network cannot encode wider contextual information in the local features, and the dependencies between different channels cannot be fully utilized. This project introduces the spatial and channel attention mechanism into the 3D U-Net. Precisely segment the tumor area in CT images of early NSCLC; extract the imaging omics features of the tumor and surrounding area, and use the intra-group correlation coefficient, consistency index and Lasso regression model to reduce the dimensionality of high-dimensional features to achieve Key feature screening, comprehensively characterize the spatial heterogeneity of tumors.
Aims
Realize the prediction of tumor heterogeneity and mutations through multimodal PET/CT images before surgery, and propose relevant explainability principles, provide reliable guidance for surgery, and promote the development of personalized medicine.
Collaborators
Mingwei Chen Guilin University Of Electronic Technology
Xipeng Pan Guilin University Of Electronic Technology