Automatic Pathological Diagnosis System for Lung Cancer Using Deep Learning
Principal Investigator
Name
Shuhao Wang
Degrees
Ph.D.
Institution
Tsinghua University
Position Title
Postdoc
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-482
Initial CDAS Request Approval
Feb 19, 2019
Title
Automatic Pathological Diagnosis System for Lung Cancer Using Deep Learning
Summary
Full digitalization of the microscopic evaluation of stained tissue sections in histopathology has become feasible in recent years, leading to the possibility for computer-aided diagnostics. In the past ten years, researchers have proposed medical diagnosis systems using deep learning. Deep learning is thoroughly studied in the field of object detection and semantic segmentation. Different from the traditional machine learning approaches, deep learning can directly learn from the medical images. It avoids the feature engineering procedure and learns key features during the model training process automatically. In this project, we will use cutting-edge deep learning approaches including to detect lung cancers in pathological slides.
Aims
1) To verify the effectiveness of different deep learning models on lung cancer prediction.
2) By carefully examining the false predictions made by the deep learning model and the learned features, we want to compare the deep learning model with human beings.
3) Due to different staining methods across hospitals, it is also interesting the performance of the deep learning model on pathological slides from various hospitals.
Collaborators
Zhuo Sun, Thorough Images
Calvin Ku, Thorough Images
Cancheng Liu, Thorough Images
Wei Xu, Tsinghua University