Effect of lung nodule size threshold on false positive CT screening rates
Principal Investigator
Name
David Gierada
Institution
Washington University
Position Title
Professor of Radiology
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-9
Initial CDAS Request Approval
Jan 23, 2013
Title
Effect of lung nodule size threshold on false positive CT screening rates
Summary
In the NLST, more than 25% of positive lung cancer CT screening tests did not lead to a diagnosis of lung cancer. These false positive screens are primarily due to the detection of small, noncalcified lung nodules, the nature of which cannot be established by a single imaging test. This leads to unnecessary additional diagnostic testing, most commonly serial CT imaging to assess for malignant growth patterns, at considerable financial expense with resultant patient anxiety and radiation exposure.
The NLST defined a positive lung cancer CT screening test by the presence of at least one noncalcified lung nodule with a maximum transverse dimension of 4 mm or larger. Most of the screen-detected nodules were near this size, and were only rarely malignant. Thus, it may be possible to reduce false positive rates by increasing the size threshold for a positive screen, with little adverse effect on lung cancer outcomes.
In this study, we will investigate the relationship between the size of lung nodules detected at CT screening and the frequency of cancer. We will thereby determine the false positive rates that would have been observed using different nodule sizes to define a screen as positive. We also will estimate the concomitant increases in false negative results (missed lung cancer) that would have occurred by raising the size threshold for positive screens. Finally, to help estimate the potential effects on outcome of a delay in diagnosis due to raising the size threshold, we will compare the stage distributions and mortality rates according to the size of the largest screen-detected nodule.
The NLST defined a positive lung cancer CT screening test by the presence of at least one noncalcified lung nodule with a maximum transverse dimension of 4 mm or larger. Most of the screen-detected nodules were near this size, and were only rarely malignant. Thus, it may be possible to reduce false positive rates by increasing the size threshold for a positive screen, with little adverse effect on lung cancer outcomes.
In this study, we will investigate the relationship between the size of lung nodules detected at CT screening and the frequency of cancer. We will thereby determine the false positive rates that would have been observed using different nodule sizes to define a screen as positive. We also will estimate the concomitant increases in false negative results (missed lung cancer) that would have occurred by raising the size threshold for positive screens. Finally, to help estimate the potential effects on outcome of a delay in diagnosis due to raising the size threshold, we will compare the stage distributions and mortality rates according to the size of the largest screen-detected nodule.
Aims
1. The determine the expected changes in false positive and false negative CT screening rates as the nodule size threshold for a positive CT screening examination is raised.
2. To estimate the effect of an increased nodule size threshold on lung cancer stage and mortality.
Collaborators
Deni Aberle, University of California at Los Angeles
Caroline Chiles, Wake Forest University
Fenghai Duan, Brown University
Bobby Nath, University of Alabama-Birmingham
Paul Pinsky, National Cancer Institute
Related Publications
-
Projected outcomes using different nodule sizes to define a positive CT lung cancer screening examination.
Gierada DS, Pinsky P, Nath H, Chiles C, Duan F, Aberle DR
J. Natl. Cancer Inst. 2014 Nov; Volume 106 (Issue 11): Pages [Epub ahead of print] PUBMED -
ROC curves for low-dose CT in the National Lung Screening Trial.
Pinsky PF, Gierada DS, Nath H, Kazerooni EA, Amorosa J
J Med Screen. 2013; Volume 20 (Issue 3): Pages 165-8 PUBMED