HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Summers, R. M. Articles by Johnson, C. D. Search for Related Content PubMed PubMed Citation Articles by Summers, R. M. Articles by Johnson, C. D.

Colonic Polyps: Complementary Role of Computer-aided Detection in CT Colonography¹

¹ From the Department of Diagnostic Radiology, Warren Grant Magnuson Clinical Center, National Institutes of Health, 10 Center Dr, MSC 1182, Bldg 10, Rm 1C660, Bethesda, MD 20892-1182 (R.M.S., A.K.J., M.F., J.D.M.); and Department of Radiology, Mayo Clinic, Rochester, Minn (C.D.J.). Received October 2, 2001; revision requested December 18; revision received February 4, 2002; accepted March 14. Supported by the intramural research programs of the Department of Diagnostic Radiology, Warren Grant Magnuson Clinical Center. Supported in part by NIH grant RO1CA75333. Address correspondence to R.M.S. (e-mail: rms@nih.gov).

View larger version (121K): [in a new window]   Figure 1a. Images of a flat, sessile, tubular 2-cm adenoma (arrows) with high-grade dysplasia that was detected in the cecum of a 75-year-old man with our CAD algorithm but not by either radiologist. (a) Image from conventional colonoscopy. (b) Perspective endoluminal CT colonographic image. The raised portion of the mass is visible. (c, d) Transverse CT colonographic images obtained with the patient in the supine position (c) without and (d) with automated detection marks. In d, an automated detection mark (small white line on edge of mass) identifies one portion of the polyp. A substantial amount of retained solid stool can be observed.

View larger version (86K): [in a new window]   Figure 1b. Images of a flat, sessile, tubular 2-cm adenoma (arrows) with high-grade dysplasia that was detected in the cecum of a 75-year-old man with our CAD algorithm but not by either radiologist. (a) Image from conventional colonoscopy. (b) Perspective endoluminal CT colonographic image. The raised portion of the mass is visible. (c, d) Transverse CT colonographic images obtained with the patient in the supine position (c) without and (d) with automated detection marks. In d, an automated detection mark (small white line on edge of mass) identifies one portion of the polyp. A substantial amount of retained solid stool can be observed.

View larger version (106K): [in a new window]   Figure 1c. Images of a flat, sessile, tubular 2-cm adenoma (arrows) with high-grade dysplasia that was detected in the cecum of a 75-year-old man with our CAD algorithm but not by either radiologist. (a) Image from conventional colonoscopy. (b) Perspective endoluminal CT colonographic image. The raised portion of the mass is visible. (c, d) Transverse CT colonographic images obtained with the patient in the supine position (c) without and (d) with automated detection marks. In d, an automated detection mark (small white line on edge of mass) identifies one portion of the polyp. A substantial amount of retained solid stool can be observed.

View larger version (107K): [in a new window]   Figure 1d. Images of a flat, sessile, tubular 2-cm adenoma (arrows) with high-grade dysplasia that was detected in the cecum of a 75-year-old man with our CAD algorithm but not by either radiologist. (a) Image from conventional colonoscopy. (b) Perspective endoluminal CT colonographic image. The raised portion of the mass is visible. (c, d) Transverse CT colonographic images obtained with the patient in the supine position (c) without and (d) with automated detection marks. In d, an automated detection mark (small white line on edge of mass) identifies one portion of the polyp. A substantial amount of retained solid stool can be observed.

View larger version (110K): [in a new window]   Figure 2a. Images of a 3.5-cm villous adenoma, detected with our CAD algorithm but not by either radiologist, in the rectum of a 54-year-old man. (Large white arrows indicate a soft-tissue mass; small white arrows indicate an automated detection mark.) (a) Image from conventional colonoscopy. (b) Perspective endoluminal CT colonographic image. (c, d) Transverse CT colonographic images of two adjacent sections obtained with the patient in the supine position. In d, the automated detection mark (small white line on edge of mass) appears small but is larger on adjacent section (c) of the top of the polyp. The tip of the rectal tube (black arrows) and false-positive locations (arrowheads in c) on rectal folds were also marked by the computer algorithm.

View larger version (89K): [in a new window]   Figure 2b. Images of a 3.5-cm villous adenoma, detected with our CAD algorithm but not by either radiologist, in the rectum of a 54-year-old man. (Large white arrows indicate a soft-tissue mass; small white arrows indicate an automated detection mark.) (a) Image from conventional colonoscopy. (b) Perspective endoluminal CT colonographic image. (c, d) Transverse CT colonographic images of two adjacent sections obtained with the patient in the supine position. In d, the automated detection mark (small white line on edge of mass) appears small but is larger on adjacent section (c) of the top of the polyp. The tip of the rectal tube (black arrows) and false-positive locations (arrowheads in c) on rectal folds were also marked by the computer algorithm.

View larger version (100K): [in a new window]   Figure 2c. Images of a 3.5-cm villous adenoma, detected with our CAD algorithm but not by either radiologist, in the rectum of a 54-year-old man. (Large white arrows indicate a soft-tissue mass; small white arrows indicate an automated detection mark.) (a) Image from conventional colonoscopy. (b) Perspective endoluminal CT colonographic image. (c, d) Transverse CT colonographic images of two adjacent sections obtained with the patient in the supine position. In d, the automated detection mark (small white line on edge of mass) appears small but is larger on adjacent section (c) of the top of the polyp. The tip of the rectal tube (black arrows) and false-positive locations (arrowheads in c) on rectal folds were also marked by the computer algorithm.

View larger version (101K): [in a new window]   Figure 2d. Images of a 3.5-cm villous adenoma, detected with our CAD algorithm but not by either radiologist, in the rectum of a 54-year-old man. (Large white arrows indicate a soft-tissue mass; small white arrows indicate an automated detection mark.) (a) Image from conventional colonoscopy. (b) Perspective endoluminal CT colonographic image. (c, d) Transverse CT colonographic images of two adjacent sections obtained with the patient in the supine position. In d, the automated detection mark (small white line on edge of mass) appears small but is larger on adjacent section (c) of the top of the polyp. The tip of the rectal tube (black arrows) and false-positive locations (arrowheads in c) on rectal folds were also marked by the computer algorithm.

View larger version (96K): [in a new window]   Figure 3a. Images of a small (described as "diminutive" by the colonoscopist) tubular adenoma (arrows) that was detected in the hepatic flexure of a 66-year-old man by our CAD algorithm but not by one of two radiologists. (a) Image from conventional colonoscopy. (b) Perspective endoluminal CT colonographic image obtained with the patient in the prone position. (c) Unmarked and (d) marked transverse CT colonographic images obtained with the patient in the prone position. A false-positive detection mark (arrowhead in d) is also seen.

View larger version (71K): [in a new window]   Figure 3b. Images of a small (described as "diminutive" by the colonoscopist) tubular adenoma (arrows) that was detected in the hepatic flexure of a 66-year-old man by our CAD algorithm but not by one of two radiologists. (a) Image from conventional colonoscopy. (b) Perspective endoluminal CT colonographic image obtained with the patient in the prone position. (c) Unmarked and (d) marked transverse CT colonographic images obtained with the patient in the prone position. A false-positive detection mark (arrowhead in d) is also seen.

View larger version (134K): [in a new window]   Figure 3c. Images of a small (described as "diminutive" by the colonoscopist) tubular adenoma (arrows) that was detected in the hepatic flexure of a 66-year-old man by our CAD algorithm but not by one of two radiologists. (a) Image from conventional colonoscopy. (b) Perspective endoluminal CT colonographic image obtained with the patient in the prone position. (c) Unmarked and (d) marked transverse CT colonographic images obtained with the patient in the prone position. A false-positive detection mark (arrowhead in d) is also seen.

View larger version (135K): [in a new window]   Figure 3d. Images of a small (described as "diminutive" by the colonoscopist) tubular adenoma (arrows) that was detected in the hepatic flexure of a 66-year-old man by our CAD algorithm but not by one of two radiologists. (a) Image from conventional colonoscopy. (b) Perspective endoluminal CT colonographic image obtained with the patient in the prone position. (c) Unmarked and (d) marked transverse CT colonographic images obtained with the patient in the prone position. A false-positive detection mark (arrowhead in d) is also seen.

View larger version (24K): [in a new window]   Figure 4. FROC curve for the computer-aided polyp detection algorithm used in this study shows results that correspond to a sensitivity value of 48% (15 false-positive detections per patient) (). The maximum achievable sensitivity for the CAD algorithm was 72% because 28% of the large polyps revealed at colonoscopy could not be located in retrospect on the CT colonographic images. The FROC curve illustrates the trade-off between sensitivity and false-positive detections. TPF = true-positive fraction per polyp. FP/patient = number of false-positive detections per patient.