Effects of In-Plane Spatial Resolution on Computer-Aided Diagnosis Features of Small Pulmonary Nodules

Principal Investigator

Name
David Gierada

Degrees
MD

Institution
Washington University

Position Title
Professor of Radiology

Email
gieradad@mir.wustl.edu

About this CDAS Project

Study
NLST (Learn more about this study)

Project ID
200909-0010

Initial CDAS Request Approval
Sep 3, 2009

Title
Effects of In-Plane Spatial Resolution on Computer-Aided Diagnosis Features of Small Pulmonary Nodules

Summary
The high prevalence of small, usually benign but indeterminate pulmonary nodules limits the specificity of CT screening for lung cancer. Analysis of pulmonary nodules using computer-aided diagnosis (CAD) methods has been shown to produce small but statistically significant improvement in the ability of radiologists to distinguish benign from malignant lesions. Studies that have investigated the added value of CAD methods have used thin section CT images, reconstructed to display the entire thorax in cross-section. It is possible, however, to increase the in-plane spatial resolution by reconstructing a complete 512 x 512 pixel CT image from a much smaller cross-sectional area. The greater detail obtained with increased in-plane spatial resolution may provide additional information for CAD helpful in further improving the distinction of benign and malignant lesions. In this study, we will explore the impact of increasing the in-plane spatial resolution on the CAD analysis of small pulmonary nodules by comparing quantitative CAD features of nodules on images reconstructed at multiple degrees of increasing in-plane spatial resolution.|

Aims

1. Compare the computer-aided diagnosis (CAD) features of screen-detected pulmonary nodules extracted from high in-plane spatial resolution images with those extracted from standard images. 2. Compare visual-ratings of screen-detected pulmonary nodule morphology evaluated on high in-plane spatial resolution images to morphology evaluated on standard images.