Narrow-band imaging comparable to white light colonoscopy in differentiating colorectal polyps

Dec 15 2008

Researchers from Yale University School of Medicine compared narrow-band imaging (NBI) without high magnification to standard white light colonoscopy in differentiating colorectal polyps during real-time colonoscopy and found that NBI was not more accurate than white light colonoscopy.

The study found, however, a significant learning curve for these experienced endoscopists using NBI as a new diagnostic tool. Once this learning curve was achieved, NBI performed significantly better. The study appears in the December issue of GIE: Gastrointestinal Endoscopy , the monthly peer-reviewed scientific journal of the American Society for Gastrointestinal Endoscopy (ASGE).

Endoscopy is a procedure that uses an endoscope -- a thin, flexible tube with a light and a lens on the end to look into the esophagus, stomach, duodenum, small intestine, colon, or rectum, in order to diagnose or treat a condition. There are many types of endoscopy, including colonoscopy, sigmoidoscopy, gastroscopy, enteroscopy, and esophogogastroduodenoscopy (EGD). Colonoscopy is frequently used to screen for colorectal cancer, which most typically develops from polyps. The two main polyp types are adenomatous, which are precancerous and hence require removal, and nonadenomatous, which do not have a high likelihood of developing into cancer.

Narrow-band imaging is a new optical technology that modifies white light by using only certain wavelengths to enhance the image. The technology provides more visual detail of the lining of the colon (including polyps) and of the small blood vessels near the surface of the polyps. At the present time, there is no widely adopted, easily applied method for distinguishing between adenomatous (potentially pre-cancerous polyps) and nonadenomatous polyps during real-time colonoscopy. From a practical standpoint, the ability to distinguish the two would allow the endoscopist to remove only those polyps with precancerous potential.

"In this prospective trial, we aimed to compare standard broadband white light colonoscopy with narrow-band imaging for the differentiation of colorectal polyps during real-time colonoscopy by using a modified Kudo pit pattern classification and vascular color intensity grading," said study lead author Jason Rogart, MD, Yale University School of Medicine, New Haven, Conn. "Overall, NBI accuracy was 80 percent, compared with 77 percent for white light alone. We also demonstrated that improved performance can be achieved over time. NBI was most useful in correctly identifying adenomas, polyps that have the most risk of developing into cancer, especially those that are small." The authors went on to state that this is, to their knowledge, the largest study to date that addresses the question of whether NBI outperforms ordinary white light in predicting the histopathologic diagnosis of polyps during real-time colonoscopy.

A total of 302 patients were enrolled from August 2006 to July 2007 at Yale University; 265 polyps were found in 131 patients resulting in an adenoma detection rate of 30 percent. Of the polyps, 49 percent were adenomas or carcinomas, whereas 51 percent were nonadenomatous; 74 percent of adenomas were 5 mm or smaller, and 42 percent were 3 mm or smaller.

Four experienced endoscopists, with a minimum number of 1,000 colonoscopies previously performed (range 1,000-10,000), participated in this study. All colonoscopies were performed with high definition endoscopes. The participating physicians were oriented to NBI before enrollment through a one-hour interactive lecture on NBI, instruction in classifying polyps based on the surface characteristics (simplified Kudo pit pattern and vascular color intensity grading), and a pretest that consisted of 20 unknown polyps photographed with the NBI system. Additionally, an atlas of endoscopic images of polyps examined with both chromoendoscopy and NBI were posted in the procedure areas.

Throughout the study period, endoscopists received feedback every two weeks about the accuracy of their endoscopic predictions compared with the histopathologic diagnosis. After enrollment was completed, the same endoscopists completed a posttest that involved the same 20 unknown polyps, which were randomly reordered. Before initiating the study, researchers planned to assess endoscopists' individual and group accuracy in the first and second halves of the study. The halfway point in the study was determined after completion of study enrollment by dividing in half the total number of polyps removed by each endoscopist.

Overall, NBI accuracy was 80 percent compared with 77 percent for white light alone. NBI performed significantly better than white light in diagnosing adenomas (sensitivity 80 percent vs. 69 percent), particularly for adenomas ¡Ü 5 mm (75 percent vs. 60 percent). There was no difference between NBI and white light for nonadenomatous polyps. Diagnostic accuracies were better for larger polyps (mean size of correct prediction 4.7 mm vs. 3.9 mm) and nonsignificant for polypoid shape (87 percent vs. 79 percent for polyps with sessile shape). Compared with white light, however, NBI did not significantly improve accuracy in any size or shape category, nor for any segment of the colon.

An equal number of polyps were analyzed in each of the two study periods (133 and 132, respectively). NBI accuracies significantly improved from 74 percent to 87 percent between the two study periods whereas, white light accuracies were unchanged (78 percent first half and 79 percent second half). In the second half of the study (i.e., after the learning curve was reached), therefore, NBI was significantly more accurate than white light.

Researchers demonstrated that there is a learning curve with regard to NBI assessment of colorectal polyps, and that NBI outperforms ordinary broadband white light once this ''learning'' is achieved. They conclude that their findings also highlight the variable NBI appearance of both adenomatous and nonadenomatous polyps. At the present time, current NBI accuracy rates of 80 percent are inadequate to defer polypectomy (polyp removal) and, therefore, limit the utility of NBI in evaluating colorectal polyps during routine clinical practice. Accuracy rates that approach 100 percent would be required for endoscopists to use this assessment to determine the need for polyp removal with confidence. Further investigation into superficial mucosal patterns of polyps and the optimal method of viewing them with NBI is needed.