Automatic Abnormality Detection and Comparison between Low Dose and Regular CT

Principal Investigator

Name
Ehsan Dehghan

Degrees
Ph.D.

Institution
IBM Research

Position Title
Research Staff Member

Email
edehgha@us.ibm.com

About this CDAS Project

Study
NLST (Learn more about this study)

Project ID
NLST-184

Initial CDAS Request Approval
Dec 10, 2015

Title
Automatic Abnormality Detection and Comparison between Low Dose and Regular CT

Summary
Automatic detection of lung abnormalities has been an active topic of research. Most of the abnormality detection algorithms were designed and trained using regular dose CT volumes. Since low-dose CT (LDCT) screening was shown to be effective in reducing the mortality due to lung cancer, the number of low-dose CT volumes is going to increase significantly. Therefore, there is a need to investigate the performance of previous abnormality detection algorithms on LDCT and develop abnormality detection algorithms for LDCT if needed.

This project will compare the performance of available lung abnormality detection algorithms on LDCT and regular CT in terms of sensitivity and specificity. New algorithms will be developed if the performance of the available algorithms on LDCT are not satisfactory for clinical use.

Aims

1- Investigating the performance of available abnormality detection algorithms on LDCT in terms of sensitivity and specificity.
2- Develop a framework for detection of lung abnormalities in LDCT.

Collaborators

Tanveer Syeda-Mahmood, IBM Research
David Beymer. IBM Research