Mar 11 2008
Novel techniques to improve the quality of imaging are helping doctors get a better look at esophageal conditions such as Barrett's esophagus (BE), but they do not necessarily improve the diagnosis or physician agreement on images, according to a study published this month in Gastroenterology, the official journal of the American Gastroenterological Association (AGA) Institute.
Researchers from the study found that the image quality of magnified images of BE that are obtained with new imaging techniques was preferred both by expert and by non-expert endoscopists. However, this higher preference did not lead to more agreement on features of interest and did not improve their ability to diagnose esophageal abnormalities.
“We know that novel technologies are dramatically improving image quality, which in turn has improved the diagnostic potential of minimally invasive techniques. But we need to review these techniques to fully understand if they deliver a clinical benefit rather than just a clearer image,” said Jacques Bergman, MD, PhD, from the Academic Medical Center in the Netherlands and lead author of the study. “In our opinion, the subjective improvement of image quality with enhancement techniques may have limited clinical relevance in this particular setting.”
In this study, the research team compared images obtained with magnifying high-resolution white light endoscopy (WLE, a standard endoscopic process), narrow band imaging (NBI), indigo carmine chromoendoscopy (ICC) and acetic acid chromoendoscopy (AAC) to determine the best techniques for detailed evaluation of BE characteristics. The team evaluated the improved quality of the images and sought to determine if the addition of these enhancement techniques would improve the agreement between observers for distinct features of disease as well as their ability to correctly identify abnormal areas. Because dysplastic and non-dysplastic BE features are associated with different visual characteristics (mucosal and vascular patterns and/or the presence of abnormal blood vessels), the type of imaging may offer improved visual diagnosis of abnormalities in the esophageal mucosal tissue.
When the enhancement techniques were compared with regular WLE images, the observers clearly preferred the enhancement techniques, which scored higher in terms of overall image quality, the quality of mucosal imaging, and the quality of vascular imaging. NBI and AAC were the most often preferred techniques and were rated best for overall image quality (43 and 40 percent of comparisons, respectively). WLE and ICC were ranked most often as the worst techniques (38 and 39 percent of comparisons, respectively). All three enhancement techniques (NBI, ICC or AAC) scored better than WLE for most image quality characteristics, but AAC rated the highest for mucosal image quality (80 percent of evaluations scored better than WLE), while NBI rated the highest for vascular image quality (60 percent of evaluations scored better than WLE).
However, the study found that the higher image quality did not translate into a clinically relevant benefit. The observer agreement was moderate on the three factors assessed on the WLE images (mucosal irregularity, vascular irregularity or the presence of abnormal blood vessels), and when WLE was combined with NBI, ICC or AAC, observer agreement did not improve. More so, expert and non-expert endoscopists did not agree more on their interpretation even when they compared all images obtained with additive techniques to WLE images of the same area.
Adding enhancement techniques also did not improve the diagnosis of neoplasia in the study. The yield for identifying early neoplasia with white light images was 86 percent for all observers, 90 percent for experts and 84 percent for non-experts, and when comparing all images, yield ranged from 70 to 86 percent for all observers. The researchers speculated on the possibility that most of the essential information may already have been present in the high-quality WLE images, or that the interpretation of the images may vary widely amongst endoscopists, which may explain the finding that expert endoscopists had a lower agreement for mucosal patterns compared to non-experts (ê = 0.43 vs. ê = 0.64).
Participants in the study were undergoing either surveillance endoscopies of known BE or work-ups of high-grade intraepithelial neoplasia (HGIN, a precancerous condition) or early cancer (EC). Conditions diagnosed in study patients included eight areas with HGIN/EC, one with low grade dysplasia, one with indefinite for dysplasia and 12 areas with non-dysplastic BE. After each patient received endoscopic imaging using WLE and each additive technique, the highest quality image from each technique was selected for analysis. The images were evaluated by seven endoscopists with no specific expertise in BE or advanced endoscopic imaging techniques and five endoscopists with experience in advanced imaging techniques and in BE to compare expert versus non-expert opinions.
Chromoendoscopy is a technique in which staining agents are applied to the gastrointestinal tract to improve the images of the esophageal mucosa captured during endoscopies. The stains enable technicians to see surface patterns or functional characteristics that may suggest early neoplasia, lesions that are difficult to detect with standard endoscopy. Recently, a new optical technique, NBI, was developed that uses light that penetrates the tissue at varying levels depending on the chosen wavelength. Light of a short wavelength (e.g. blue light in the visible spectrum) penetrates very little, allowing for improved surface detail imaging. Two other additives were applied to chromoendoscopy for this study – ICC and AAC.