Accuracy of the Differential Diagnosis of Colorectal Serrated Polyps Using a Conventional Endoscope: A Prospective Study

抄録

Some serrated polyps （SPs） are thought to be precursors of colorectal cancers. However, the endoscopic diagnosis of sessile serrated adenoma/polyps （SSA/Ps） has been reported to have a low accuracy. The aim of this study was to clarify the ability to distinguish between SSA/Ps and non-SSA/Ps by using mucosal crypt patterns combined with endoscopic findings. In total, 457 consecutive patients who underwent endoscopic resection for colorectal polyps at the Showa University Hospital from April 2007 to December 2010 were prospectively enrolled in this study. Before treatment, mucosal crypt patterns of the lesions were classified into three types （hyperplastic, adenomatous, and mixed pattern）. When the lesion had an adenomatous pattern with a cerebriform appearance or mixed pattern, it was diagnosed as a traditional serrated adenoma （TSA）. If the lesion had a hyperplastic pattern and was sized 6mm or more in the proximal colon or 10mm or more in the distal colon, it was diagnosed as an SSA/P by the endoscopist. We analyzed 1,151 colorectal polyps in this study. Endoscopically, 117 polyps were diagnosed as SSA/Ps or hyperplastic polyps （HPs）, 998 polyps were conventional adenomas, and 36 polyps were TSAs, with diagnostic accuracies of 94.7％, 94.1％, and 97.3％, respectively. Of the 117 polyps diagnosed as SSA/Ps or HPs, 59 lesions met our criteria for SSA/Ps, with a diagnostic accuracy of 70.9％. Our results indicate that the combination of mucosal crypt patterns and endoscopic findings may be useful for differentiating between SPs and non-SPs. However, additional specific endoscopic features of SSA/Ps are still needed.