| Theme |
Image-enhanced endoscopy -- Current status and significance of colorectal NBI and BLI |
| Title |
Efficacy of and problems with Capillary pattern classification using NBI for diagnosis of colorectal lesions |
| Author |
Hiroaki Ikematsu |
Division of Endoscopy & Gastrointestinal Oncology, National Cancer Center Hospital East |
| Author |
Tomoji Kato |
Division of Endoscopy & Gastrointestinal Oncology, National Cancer Center Hospital East |
| Author |
Shozo Osera |
Division of Endoscopy & Gastrointestinal Oncology, National Cancer Center Hospital East |
| Author |
Hideaki Miyamoto |
Division of Endoscopy & Gastrointestinal Oncology, National Cancer Center Hospital East |
| Author |
Yusuke Yoda |
Division of Endoscopy & Gastrointestinal Oncology, National Cancer Center Hospital East |
| Author |
Tomoyuki Odagaki |
Division of Endoscopy & Gastrointestinal Oncology, National Cancer Center Hospital East |
| Author |
Yasuhiro Oono |
Division of Endoscopy & Gastrointestinal Oncology, National Cancer Center Hospital East |
| Author |
Kazuhiro Kaneko |
Division of Endoscopy & Gastrointestinal Oncology, National Cancer Center Hospital East |
| Author |
Yasushi Sano |
Sano Hospital |
| [ Summary ] |
Observation of micro-capillary (MC) vessels using Narrow Band Imaging (NBI) which requires no application of dye solutions has recently been used for colorectal diagnosis. A problem exists in that there are many NBI classifications for colorectal lesions. Our aim was to clarify the diagnostic accuracy of and problems with the Capillary pattern system for NBI classification now in use. Lesions evaluated by both pit patterns and Capillary patterns were derived from our database. Diagnostic accuracy, sensitivity, and specificity were analyzed on the basis of individual pathological reports after endoscopic removal or surgery. From October 2005 to March 2009, 1,731 eligible lesions were examined in this study. The diagnostic accuracy, sensitivity, and specificity to distinguish between non-neoplastic and neoplastic lesions were 98.5 %, 99.0 %, and 92.4 %, equivalent to those obtained through pit pattern classification. The diagnostic accuracy, sensitivity, specificity, negative predictive values (NPV) and positive predictive values (PPV) for depth infiltration of neopastic lesions were 97.7 %, 73.2 %, 98.9 %, 98.7 % and 75.3 %. Our results regarding depth of infiltration, especially in PPV, were somewhat inferior to those obtained for evaluation using pit pattern classifications. NBI was the most reliable method to distinguish between non-neoplastic and neoplastic lesions. In contrast, diagnostic ability when using NBI magnifying observation for lesions infiltrating into deeper submucosal layers was inferior to the diagnostic ability for use of pit patterns with dye solutions. Furthermore, there was a problem with formulating diagnoses because depressed lesions present various vascular findings. |