Su1553 The Impact of Vascular and Surface Patterns in the Differentiation Diagnosis of Diminutive Colorectal Polyps Using NBI With Magnification

BackgroundMeshed capillary vessels (MCV) and pit (Pit) pattern seen on the colonic polyp surface using narrow-band imaging with magnification (NBIM) has been reported to be able to help differentiate between adenomatous and non-adenomatous polyps. We hypothesized that the efficacy of NBIM might differ based on the application of its classification or experience of each endoscopist.AimsThe aim of this study was to assess the diagnostic accuracy of MCV and Pit using NBIM for differentiation of the colorectal polyps according to the experience of endoscopists.Patients and methodsThis study prospectively enrolled consecutive patients who underwent magnification colonoscopy (CF-H260AZI and PCF-240ZI, Olympus Co., Tokyo, Japan) between September and December 2008 at our hospital. A total of 52 colorectal polyps ≤ 5 mm from 43 patients were observed with NBIM and then the polyps were removed for histopathological analysis by EMR, polypectomy or biopsy procedure. A total of 104 NBIM images of the polyp MCV and Pit patterns obtained in our trial were retrieved. These endoscopic images were reviewed in randomized order by five endoscopists; one highly experienced endoscopist (HEE) with over five years of NBIM experience and four less experienced endoscopists (LEEs) with less than one year to predict polyp histology based on MCV and Pit patterns. Those patterns were assessed using NBIM for MCV; invisible or faintly visible MCV as a non-neoplastic and clearly visible MCV as a neoplastic and Kudo pit pattern; III-V as a neoplastic, I-II as a non-neoplastic, respectively. Data was compared with histopathology as the gold standard and cross tabulations of MCV and Pit yielded four groups; lesions were diagnosed as neoplastic by MCV alone(A), Pit alone(B), combining both MCV and Pit(C) and either MCV, Pit or both(D).ResultsThe mean polyp size was 3.3mm (range 1 - 5mm). Of the 52 colorectal polyps, 33 polyps were adenomas, 19 polyps were hyperplastic polyps. Accuracy/sensitivity/specificity for HEE were A: 92%/100%/79%, B: 69%/58%/89% C: 69%/58%/89%, D: 92%/100%/79%, respectively. For LEE, A: 87%/95%/74%, B: 79%/80%/78%, C: 80%/80%/79%, D: 87%/95%/72%, respectively. MCV had significantly demonstrated higher accuracy and sensitivity than Pit for both HEE and LEEs. There were no additional benefit combinations of MCV and Pit compared to MCV alone; for HEE, A vs. B; p=0.0041/p<0.001/p=0.5, A vs. C; p=0.0041/p<0.001/p=0.5, A vs. D; p=1.0/p=1.0/p=1.0; for LEEs, A vs. B; p=0.0022/p<0.001/p=0.38, A vs. C; p=0.0035/p<0.001/p=0.13, A vs. D; p=1.0/p=1.0/p=1.0.ConclusionMCV and Pit patterns using MBIM were both highly accurate in characterizing neoplastic colonic polyps, however, NBIM by using MCV alone was considered to be reasonable for differentiating diminutive colonic polyps in both HEE and LEEs. BackgroundMeshed capillary vessels (MCV) and pit (Pit) pattern seen on the colonic polyp surface using narrow-band imaging with magnification (NBIM) has been reported to be able to help differentiate between adenomatous and non-adenomatous polyps. We hypothesized that the efficacy of NBIM might differ based on the application of its classification or experience of each endoscopist. Meshed capillary vessels (MCV) and pit (Pit) pattern seen on the colonic polyp surface using narrow-band imaging with magnification (NBIM) has been reported to be able to help differentiate between adenomatous and non-adenomatous polyps. We hypothesized that the efficacy of NBIM might differ based on the application of its classification or experience of each endoscopist. AimsThe aim of this study was to assess the diagnostic accuracy of MCV and Pit using NBIM for differentiation of the colorectal polyps according to the experience of endoscopists. The aim of this study was to assess the diagnostic accuracy of MCV and Pit using NBIM for differentiation of the colorectal polyps according to the experience of endoscopists. Patients and methodsThis study prospectively enrolled consecutive patients who underwent magnification colonoscopy (CF-H260AZI and PCF-240ZI, Olympus Co., Tokyo, Japan) between September and December 2008 at our hospital. A total of 52 colorectal polyps ≤ 5 mm from 43 patients were observed with NBIM and then the polyps were removed for histopathological analysis by EMR, polypectomy or biopsy procedure. A total of 104 NBIM images of the polyp MCV and Pit patterns obtained in our trial were retrieved. These endoscopic images were reviewed in randomized order by five endoscopists; one highly experienced endoscopist (HEE) with over five years of NBIM experience and four less experienced endoscopists (LEEs) with less than one year to predict polyp histology based on MCV and Pit patterns. Those patterns were assessed using NBIM for MCV; invisible or faintly visible MCV as a non-neoplastic and clearly visible MCV as a neoplastic and Kudo pit pattern; III-V as a neoplastic, I-II as a non-neoplastic, respectively. Data was compared with histopathology as the gold standard and cross tabulations of MCV and Pit yielded four groups; lesions were diagnosed as neoplastic by MCV alone(A), Pit alone(B), combining both MCV and Pit(C) and either MCV, Pit or both(D). This study prospectively enrolled consecutive patients who underwent magnification colonoscopy (CF-H260AZI and PCF-240ZI, Olympus Co., Tokyo, Japan) between September and December 2008 at our hospital. A total of 52 colorectal polyps ≤ 5 mm from 43 patients were observed with NBIM and then the polyps were removed for histopathological analysis by EMR, polypectomy or biopsy procedure. A total of 104 NBIM images of the polyp MCV and Pit patterns obtained in our trial were retrieved. These endoscopic images were reviewed in randomized order by five endoscopists; one highly experienced endoscopist (HEE) with over five years of NBIM experience and four less experienced endoscopists (LEEs) with less than one year to predict polyp histology based on MCV and Pit patterns. Those patterns were assessed using NBIM for MCV; invisible or faintly visible MCV as a non-neoplastic and clearly visible MCV as a neoplastic and Kudo pit pattern; III-V as a neoplastic, I-II as a non-neoplastic, respectively. Data was compared with histopathology as the gold standard and cross tabulations of MCV and Pit yielded four groups; lesions were diagnosed as neoplastic by MCV alone(A), Pit alone(B), combining both MCV and Pit(C) and either MCV, Pit or both(D). ResultsThe mean polyp size was 3.3mm (range 1 - 5mm). Of the 52 colorectal polyps, 33 polyps were adenomas, 19 polyps were hyperplastic polyps. Accuracy/sensitivity/specificity for HEE were A: 92%/100%/79%, B: 69%/58%/89% C: 69%/58%/89%, D: 92%/100%/79%, respectively. For LEE, A: 87%/95%/74%, B: 79%/80%/78%, C: 80%/80%/79%, D: 87%/95%/72%, respectively. MCV had significantly demonstrated higher accuracy and sensitivity than Pit for both HEE and LEEs. There were no additional benefit combinations of MCV and Pit compared to MCV alone; for HEE, A vs. B; p=0.0041/p<0.001/p=0.5, A vs. C; p=0.0041/p<0.001/p=0.5, A vs. D; p=1.0/p=1.0/p=1.0; for LEEs, A vs. B; p=0.0022/p<0.001/p=0.38, A vs. C; p=0.0035/p<0.001/p=0.13, A vs. D; p=1.0/p=1.0/p=1.0. The mean polyp size was 3.3mm (range 1 - 5mm). Of the 52 colorectal polyps, 33 polyps were adenomas, 19 polyps were hyperplastic polyps. Accuracy/sensitivity/specificity for HEE were A: 92%/100%/79%, B: 69%/58%/89% C: 69%/58%/89%, D: 92%/100%/79%, respectively. For LEE, A: 87%/95%/74%, B: 79%/80%/78%, C: 80%/80%/79%, D: 87%/95%/72%, respectively. MCV had significantly demonstrated higher accuracy and sensitivity than Pit for both HEE and LEEs. There were no additional benefit combinations of MCV and Pit compared to MCV alone; for HEE, A vs. B; p=0.0041/p<0.001/p=0.5, A vs. C; p=0.0041/p<0.001/p=0.5, A vs. D; p=1.0/p=1.0/p=1.0; for LEEs, A vs. B; p=0.0022/p<0.001/p=0.38, A vs. C; p=0.0035/p<0.001/p=0.13, A vs. D; p=1.0/p=1.0/p=1.0. ConclusionMCV and Pit patterns using MBIM were both highly accurate in characterizing neoplastic colonic polyps, however, NBIM by using MCV alone was considered to be reasonable for differentiating diminutive colonic polyps in both HEE and LEEs. MCV and Pit patterns using MBIM were both highly accurate in characterizing neoplastic colonic polyps, however, NBIM by using MCV alone was considered to be reasonable for differentiating diminutive colonic polyps in both HEE and LEEs.