DOI: 10.1148/radiol.2462071140
(Radiology 2008;246:648-649.)
© RSNA, 2008
Missed Lesions at Primary 2D CT Colonography: Further Support for 3D Polyp Detection
Perry J. Pickhardt, MD
Department of Radiology, University of Wisconsin Medical School, E3/311 Clinical Science Center, 600 Highland Avenue, Madison, WI 53792-3252
e-mail: pj.pickhardt{at}hosp.wisc.edu
Editor:
I read with interest the excellent article by Mr Doshi and colleagues (1) in the July 2007 issue of Radiology. Given the relative disparity of performance results to date, the underlying causes of false-negative interpretations at computed tomographic (CT) colonography are of particular relevance. On retrospective review of the missed lesions at primary two-dimensional (2D) CT colonography evaluation from the original trial by Rockey et al (2), the investigators (1) found that observer perceptual error was the predominate cause. Of note, failure to prospectively detect clearly identifiable lesions with primary 2D evaluation accounted for 13 (65%) of the 20 large adenomas and cancers that were missed. Interestingly, when these readily preventable 2D perceptual errors are eliminated, the sensitivity of CT colonography for large neoplasms in this trial closely matches the results obtained with primary three-dimensional (3D) endoluminal polyp detection (3–5). In our own experience with primary 3D polyp detection (supplemented by 2D evaluation), observer perceptual error rarely results in false-negative interpretations at CT colonography.
This fundamental performance difference between primary 2D and primary 3D polyp detection is further illustrated by the results from retrospective primary 2D review of CT colonography cases from the Department of Defense trial (6), which was performed by highly trained and experienced readers. In that study, sensitivity of primary 2D polyp detection at the 6- and 10-mm thresholds decreased to 37.9% and 60.4%, respectively, results which closely resemble the primary 2D results from the trials by Rockey et al (2), Cotton et al (7), and Johnson et al (8). Similar to the findings from Mr Doshi and colleagues, most 2D false-negative polyps were clearly identifiable in retrospect.
Mr Doshi and colleagues suggest that sensitivity of polyp detection could be improved primarily through the use of computer-aided detection and more rigorous training of readers. As for the potential impact of primary 3D evaluation on polyp detection sensitivity, the authors believed that this "requires further investigation." In my opinion, findings of this study only add to the now convincing body of evidence that the standard 3D endoluminal display is an easily mastered and reproducibly effective means for polyp detection at CT colonography. If this simple message were to be embraced by the CT colonography community at large, I believe that we could finally move beyond the prolonged validation phase of this effective screening tool and focus on widespread implementation.
The author is a consultant for Viatronix, Fleet, MedicSight, and Philips Medical Systems.
 |
References
|
|---|
- Doshi T, Rusinak D, Halvorsen RA, Rockey DC, Suzuki K, Dachman AH. CT colonography: false-negative interpretations. Radiology 2007;244:165–173. [Abstract/Free Full Text]
- Rockey DC, Paulsen EK, Niedzwiecki D, et al. Analysis of air contrast barium enema, computed tomographic colonography, and colonoscopy: prospective comparison. Lancet 2005;365:305–311. [Medline]
- Pickhardt PJ, Choi JR, Hwang I, et al. Computed tomographic virtual colonoscopy to screen for colorectal neoplasia in asymptomatic adults. N Engl J Med 2003;349:2191–2200. [Abstract/Free Full Text]
- Cash BD, Kim C, Cullen P, et al. Accuracy of computed tomographic colonography for colorectal cancer screening in asymptomatic individuals [abstr]. In: Digestive Disease Week 2006 annual meeting program. Los Angeles CA, May 20–25, 2006; 473.
- Graser A, Kolligs FT, Kramer H, Reiser MF, Becker C. Results from the "Munich Colorectal Cancer Prevention Trial": comparison of low-dose 64-MDCT colonography and video colonoscopy in a screening population [abstr]. In: Radiological Society of North America Scientific Assembly and Annual Meeting Program. Oak Brook, Ill: Radiological Society of North America, 2006; 356.
- Pickhardt PJ, Lee AD, Taylor AJ, et al. Primary 2D versus primary 3D polyp detection at screening CT colonography. AJR Am J Roentgenol 2007;189:1451–1456. [Abstract/Free Full Text]
- Cotton PB, Durkalski VL, Pineau BC, et al. Computed tomographic colonography (virtual colonoscopy): a multicenter comparison with standard colonoscopy for detection of colorectal neoplasia. JAMA 2004;291:1713–1719. [Abstract/Free Full Text]
- Johnson CD, Harmsen WS, Wilson LA, et al. Prospective blinded evaluation of computed tomographic colonography for screen detection of colorectal polyps. Gastroenterology 2003;125:311–319.[CrossRef][Medline]
Response
Abraham H. Dachman, MD
Department of Radiology, University of Chicago, Department of Radiology, MC 2026, 5841 S Maryland Ave, Chicago, IL 60637
e-mail: ahdachma{at}uchicago.edu
We thank Dr Pickhardt for his comments. However, logic dictates that while the statement that "standard 3D endoluminal display is an easily mastered and reproducibly effective means for polyp detection at CT colonography" may be true, we do not believe that it is possible to draw this conclusion from our data (1). To demonstrate a clear difference between these two modes of reading, more rigorous investigation is required. For example, rereading of the entire data set would have been informative but was not feasible. Indeed, as emphasized in the discussion of our article, "no attempt was made to reread the entire data set for all patients in our cohort."
We agree that improved reader performance by using primary 3D reading with 2D problem-solving is one possible inference from our data. It was therefore mentioned as such in the discussion in the context of other reading methods. In particular, 3D paradigms of interpretation are now expanding to include nonstandard views that may ultimately prove to be faster to use and easier for novices to learn to interpret. We refer the readers to a recent publication on that topic (2). That said, many of us who participated in the original trial in fact now use both full 3D and 2D reading routinely for CT colonographic interpretation.
|
References
|
|---|
- Doshi T, Rusinak D, Halvorsen RA, Rockey DC, Suzuki K, Dachman AH. CT colonography: false-negative interpretations. Radiology 2007;244:165–173. [Abstract/Free Full Text]
- Dachman AH, Lefere P, Gryspeerdt S, Morin M. CT colonography: visualization methods, interpretation, and pitfalls. Radiol Clin North Am 2007;45(2):347–359.[CrossRef][Medline]
TOP
References
References 
