DOI: 10.1148/radiol.2393050418
CT Colonography: Influence of 3D Viewing and Polyp Candidate Features on Interpretation with Computer-aided Detection1
Rong Shi, MD,
Pamela Schraedley-Desmond, PhD,
Sandy Napel, PhD,
Eric W. Olcott, MD,
R. Brooke Jeffrey, Jr, MD,
Judy Yee, MD,
Michael E. Zalis, MD,
Daniel Margolis, MD,
David S. Paik, PhD,
Anthony J. Sherbondy, MS,
Padmavathi Sundaram, MS and
Christopher F. Beaulieu, MD, PhD
1 From the Department of Radiology, Stanford University Medical Center, James H. Clark Center, 318 Campus Dr, Room S324, Stanford, CA 94305-5450 (R.S., P. Schraedley-Desmond, S.N., E.W.O., R.B.J., D.M., D.S.P., A.J.S., P. Sundaram, C.F.B.); Veterans Affairs Palo Alto Health Care System, Palo Alto, Calif (E.W.O.); Department of Radiology, University of California, San Francisco, San Francisco, Calif (J.Y.); San Francisco Veterans Affairs Medical Center, San Francisco, Calif (J.Y.); and Department of Radiology, Massachusetts General Hospital, Boston, Mass (M.E.Z.). Received March 11, 2005; revision requested May 4; revision received May 20; accepted June 20; final version accepted August 24. Supported by the National Institutes of Health (R01 CA72023 and 1 U54 GM072970) and the Lucas Foundation.
Address correspondence to R.S. (e-mail: rshi{at}stanford.edu).

View larger version (84K):
[in a new window]
|
Figure 1: Magnified, A, transverse, and, B, coronal SURF interface sections show an experimental polyp, which was confirmed with colonoscopy, in the center of the bounding box. C, Volume-rendering window, which was enabled during reference standard setting and combined 2D and 3D reading and disabled during 2D reading, shows a subvolume surrounding the polyp candidate. The polyp (arrow) and haustral fold (arrowheads) were added. D, Polyp CAD manager used to establish the reference standard in the blinded reader trial.
|
|

View larger version (63K):
[in a new window]
|
Figure 2: Polyp CAD manager window used with SURF. The left panel lists polyp candidates presented to each reader. During reference standard setting, all panels were used to characterize TP and FP polyp candidates, assign a confidence level, and rate each feature according to parameters shown. During blinded reader review, only the subpanel (indicated by the dashed box) was used to enter confidence levels for 2D or combined 2D and 3D viewing (ie, study readers were blinded to information concerning additional features and optical colonoscopy results).
|
|

View larger version (37K):
[in a new window]
|
Figure 3: Receiver operating characteristic curves for individual radiologists and lesions of all sizes. For readers 1, 3, and 4, there was a significant increase in the Az value with combined 2D and 3D viewing (2D3D) during interpretation. Reader 2 demonstrated extremely good performance with 2D viewing alone; thus, the addition of 3D viewing was not associated with a significant increase in the Az value.
|
|
Copyright © 2006 by the Radiological Society of North America.