Image processing algorithm to screen for lung cancer on radiology images

Principal Investigator

Name
Alexandros Karargyris

Degrees
PhD

Institution
IBM

Position Title
Software Engineering Researcher

Email
akararg@us.ibm.com

About this CDAS Project

Study
PLCO (Learn more about this study)

Project ID
PLCO-214

Initial CDAS Request Approval
Jun 15, 2016

Title
Image processing algorithm to screen for lung cancer on radiology images

Summary
Medical Sieve is an ambitious long-term exploratory grand challenge project to build a next generation cognitive assistant with advanced multimodal analytics, clinical knowledge and reasoning capabilities that is qualified to assist in clinical decision making in radiology and cardiology. The project aims at producing a sieve that filters essential clinical and diagnostic imaging information to form anomaly-driven summaries and recommendations that tremendously reduce the viewing load of clinicians without negatively impacting diagnosis. More information about the project can be found here: http://researcher.watson.ibm.com/researcher/view_group.php?id=4384

Aims

1) To investigate whether hardcrafted image features can help with CXR image classification as normal versus abnormal in terms of lung disease (e.g. lung cancer). To provide accuracy results.
2) To investigate whether deep learning approaches can help with CXR image classification as normal versus abnormal in terms of lung disease (e.g. lung cancer). To provide accuracy results.
3) To investigate whether patient demographics and metadata from patients' records coupled with aforementioned image features can improve overall classification accuracy.
4) To investigate whether aforementioned image features can help with disease progression prediction.

Collaborators

Tanveer Syeda-Mahmood
Hongzhi Wang
Ehsan Dehgan Marvast
Tyler Baldwin