| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Royal Hallamshire Hospital Sheffield, UK
University of Sheffield Medical School Sheffield, UK
Royal Hallamshire Hospital Sheffield, UK
In a recent AJR article, Dr. Pickhardt and colleagues [1] conclude that flat adenomas that are 6 mm or greater in diameter are uncommon in a typical Western population undergoing screening for colorectal cancer and that advanced flat neoplastic lesions are rare. Furthermore, Pickhardt et al. indicate that flat lesions are not a significant drawback for virtual colonoscopic screening. However, we believe that it is important to highlight new developments in optical colonoscopy and enhanced lesion detection methods when interpreting their data. Of note is that Pickhardt et al. performed conventional colonoscopic techniques without the use of high-magnification or chromoscopic imaging for index endoscopic colorectal assessment in a group of asymptomatic individuals who underwent screening.
Chromoscopic colonoscopy improved the detection of Paris type II lesions, which can exhibit only subtle mucosal changes, using conventional imaging. The Sheffield group [2] reported a total of 458 flat lesions in a cohort of 850 patients stratified as at moderate risk for colorectal cancer. Of the 267 Paris type II adenomatous lesions, 82% (54/66) of those having high-grade dysplasia were located in the right colon, 15% (10/66) in the left, and 3% (2/66) in the rectum. Furthermore, depressed lesions and large flat adenomas (> 8 mm in diameter) were more likely to contain high-grade dysplasia, with 54% of early carcinomas assuming a flat or type IIc component.
The detection of flat and diminutive colorectal lesions using panchromoscopic colonoscopy has now been shown in two randomized prospective trials in cohorts at moderate risk of colorectal cancer. In their study, Brooker et al. [3] found 3% of diminutive lesions showed moderate dysplastic changes at histopathology with 2% having a villous architecture [3]. Despite their diminutive appearance (< 5 mm), these lesions were not diagnosed at index "white light" conventional colonoscopy and are of clear clinical significance assuming a high risk of progression to invasive carcinoma. Additional data to corroborate these findings were shown by results of a randomized controlled trial (extubation time and lavage effect controlled) performed by the Sheffield group.
Another study using a back-to-back study design (conventional colonoscopy vs pancolonic high-magnification chromoscopic colonoscopy) in patients fulfilling modified Amsterdam criteria for hereditary nonpolyposis colorectal cancer also revealed superior detection rates [4]. Using segmental pancolonic chromoscopy, an additional 52 lesions that were not apparent using conventional colonoscopy and targeted chromoscopy (17 Paris type I and 35 Paris type II) were detected in 15 patients. The median size for Paris type I and II lesions was 7 mm (range, 2-14 mm) and 4 mm (range, 1-8 mm), respectively. Thirty-two additional adenomas in 17 patients were detected (13 Paris type I [40%] and 19 Paris type II [60%]), of which seven had foci of high-grade dysplasia (71% in Paris type II lesions).
Given these data, we suggest that first-line colorectal investigations in a screening cohort using CT virtual colonoscopy requires further clarification. A segmental, sequential unblinded study using a chromoscopic colonoscopy technique may be required to clarify the comparable diagnostic abilities of these investigatory techniques.
 Top References References |
|---|
University of Wisconsin Medical School Madison, WI 53792-3252
I thank Dr. Hurlstone and colleagues for their interest in our work [1]. The topic of flat colorectal lesions remains a potential source of controversy with regard to screening techniques. However, since the time our work appeared in the AJR [1], a key article, which draws on the invaluable data resources from the National Polyp Study, focusing on the significance of flat adenomas has been formally published [2].
The National Polyp Study is a U.S. multicenter clinical trial that was initiated in 1980 and has since provided landmark evidence regarding the benefit of screening for colorectal cancer using colonoscopy [3]. Grading of dysplasia for all adenomas harvested in this study was performed prospectively by a three-person pathology review team. Importantly, histologic grading was not biased by knowledge of flat-versus-polypoid (nonflat) morphology. This study found that flat adenomas had the lowest rate of high-grade dysplasia (1.3%) compared with rates of 7.4% and 10.0% for sessile and pedunculated polyps, respectively [2]. Villous histology and polyp size—not flat morphology—were independent risk factors for high-grade dysplasia [4]. Furthermore, patients with flat adenomas were not found to be at greater risk for advanced adenomas at subsequent surveillance colonoscopy [2]. These collective results contradict the suggestion that flat adenomas represent an important aggressive subset of colorectal neoplasia, at least not for a U.S. population.
The fact that the colonoscopies performed as part of the National Polyp Study did not use the high-magnification chromoscopic techniques advocated by Hurlstone and colleagues raises the possibility that significant flat lesions may have been missed. However, there are at least two reasons this is unlikely. First, the relative frequency of flat adenomas detected in the National Polyp Study was similar to those reported by studies for which chromoendoscopic techniques were used [2]. Second, if a significant number of aggressive flat neoplasms had been missed at initial examination, more incident cancers should have developed over the course of longitudinal evaluation. Hence, a drastic policy change to implement high-magnification chromoscopic techniques for routine optical colonoscopy screening is not justified for practice in the United States [5, 6]. By extension, the need for a new reference standard for validating virtual colonoscopy beyond our enhanced method of segmental unblinding is also not warranted.
By applying state-of-the-art virtual colonoscopy techniques, including primary 3D polyp detection and oral contrast tagging, we have found that the sensitivity for detecting flat lesions is similar to that for detecting polypoid (nonflat) lesions [1]. Combined with the fact that histologically advanced flat neoplasms, particularly those measuring less than 1 cm, seem to be quite rare in the United States, we maintain that they do not represent a significant drawback for colonoscopic screening of average-risk adults, whether by virtual or direct optical techniques. Certain high-risk populations, such as patients with hereditary nonpolyposis colorectal cancer syndromes, are better served by primary optimal colonoscopy given its abilities for tissue sampling and therapy.
Compliance with colorectal screening programs is clearly the greatest obstacle in adequately addressing this deadly but largely preventable condition. Complete structural evaluation of the colon with emphasis on detecting large polyps in the majority of the screening population represents a logical strategy. To this end, implementation of wide-spread virtual colonoscopy screening that uses proven techniques [7], in conjunction with existing optical colonoscopy screening efforts, is a most worthy endeavor that should proceed without delay.
|
TopReferencesReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  P. J. Pickhardt Screening CT Colonography: How I Do It Am. J. Roentgenol., August 1, 2007; 189(2): 290 - 298. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
