Adding narrow-band imaging and autofluorescence imaging to high-resolution endoscopy – a combination known as endoscopic trimodal imaging – doesn’t improve the detection of colonic adenomas in routine clinical practice, compared with standard video endoscopy, Dr. Teaco Kuiper and colleagues reported in the June issue of Gastroenterology.
This result is somewhat surprising. Not only does endoscopic trimodal imaging (ETMI) involve newer technology specifically designed to improve detection and differentiation of gastrointestinal lesions, but in this study it also entailed a significantly longer time for inspection of the colon, "which is known to positively influence adenoma detection rate," wrote Dr. Kuiper of the University of Amsterdam and associates (Gastroenterology 2011 June [doi:10.1053/j.gastro.2011.03.008]).
Nevertheless, a colonoscopy involving an initial pass with a high-resolution endoscope yielded nearly the same detection rate as did an initial pass with a standard video endoscope, and a second pass adding narrow-band imaging (NBI) and autofluorescence imaging (AFI) to the high-resolution endoscopy yielded nearly the same detection rate as did a second pass with a standard video endoscope, the researchers said.
They assessed the value of various endoscopic imaging techniques in routine clinical practice because until now, "nearly all studies evaluating narrow-band imaging and autofluorescence imaging have been performed in expert centers. As a consequence, these studies often include a selected patient population and are performed by a small number of endoscopists with extensive experience in advanced clinical imaging," they noted.
"There is a need to ascertain if these new imaging techniques improve detection and differentiation of colonic lesions in the community and nonspecialised settings."
Dr. Kuiper and colleagues compared ETMI with standard video endoscopy at six nonacademic medical centers in the Amsterdam region over a 3-year period. The study included unselected patients scheduled for routine surveillance colonoscopy, and all the procedures were performed by eight endoscopists with experience in regular (but not advanced) clinical imaging, with more than 2,000 colonoscopies each.
The study subjects were adults undergoing colonoscopy because of a history of adenomatous polyps, colorectal cancer for which partial colectomy had been performed, hereditary nonpolyposis colorectal cancer, or a family history of colorectal cancer. They were randomly assigned to undergo either ETMI (118 patients) or standard video endoscopy (116 patients) under conscious sedation. Only patients with adequate bowel preparation to achieve adequate imaging were included in the study.
In the ETMI group, patients underwent a first-pass inspection with high-resolution endoscopy alone followed by a second pass with autofluorescence imaging and narrow-band imaging. To make the control procedure closely resemble this approach, patients in the control group underwent a first-pass inspection with standard video endoscopy immediately followed by a second pass with the same standard video endoscopy.
The mean inspection time was significantly longer with ETMI during both the first pass (7.06 minutes) and second pass (7.34 minutes) than with standard video endoscopy (6.18 minutes and 6.06 minutes).
The primary outcome measure was the mean number of adenomas detected by the two methods. By this measure, trimodal imaging detected 1.03 adenomas and standard imaging detected 0.97, a nonsignificant difference.
A total of 87 adenomas were detected during the first pass with high-resolution endoscopy, including 46 advanced adenomas. In the control group, 79 adenomas were detected during the first pass with standard endoscopy, including 33 advanced adenomas. These rates were not significantly different.
A total of 34 adenomas were detected during the second pass with autofluorescence imaging and narrow-band imaging added to high-resolution endoscopy, including 12 that were advanced. In the control group, an additional 33 adenomas were detected during the second pass with standard endoscopy, including 14 advanced adenomas. These rates also were not significantly different.
Nine patients had adenomas that were detected only during the second pass with autofluorescence imaging and narrow-band imaging. Similarly, 9 patients in the control group had adenomas that were detected only during the second pass with standard endoscopy.
The adenoma miss rate was 28% after the second pass in the trimodal imaging group, and the miss rate of advanced adenomas was 21%. The corresponding miss rates in the control group were 29% and 30%.
These miss rates were substantial and much higher than has been reported in the literature, despite the fact that patients with poor bowel preparation were excluded and that inspection time was longer than 6 minutes in most cases. These rates are high in comparison to those in clinical trials, but may well reflect the typical rates in true day-to-day practice, Dr. Kuiper and colleagues said.
In this study, autofluorescence imaging had a sensitivity of 87%, and narrow-band imaging had a sensitivity of 90%. These rates are "clinically unacceptable, as approximately 10% of all adenomas would be left in situ," the researchers noted.