- Email: [email protected]
Su1531 Narrow Band Imaging Versus I-Scan for Real-Time Histologic Prediction of Diminutive Colonic Polyps: A Prospective Comparative Study
Abstracts
discrete lesions or flat mucosa). Seventy-five percent (33/44) had been on 5-ASA agents for a median of 9 years. Median total procedure time was 50mins. The median total examination time was identical with NBI and CE at 13mins, P⫽0.22. Overall 144 colonic lesions were detected, 21 of which were dysplastic. CE detected 131 lesions compared to 102 for NBI, P⬍0.015. There was no statistically significant difference in the number of dysplastic lesions detected by CE (20/21) and NBI (17/21). Fifty-five percent (79/144) of the lesions were 5mm or less in size. Most lesions were flat (Paris II: 53%, 76/144) or sessile (Paris Is: 36%, 52/144). The Kudo pattern was benign (I or II) in 84% (92/132) of lesions. Twenty-nine lesions (20%) had advanced Kudo patterns (IIIL, 26/144; IV, 3/144). The accuracy of the Kudo classification by NBI for neoplasia was 74% (95% CI: 0.68-0.80). In lesions occurring in areas of inflammation, Kudo accuracy for neoplasia was lower at 44% (95% CI: 0.35-0.51) compared to the rest of the lesions where accuracy was 75% (95% CI: 0.71-0.82). Summary: Chromoendoscopy identifies more discrete lesions during surveillance colonoscopy than narrow band imaging, and may be the preferred modality. Most lesions detected, however, are non-dysplastic. Most lesions are flat or sessile. The morphological Kudo classification appears to be less accurate at correlating with histology than in the non-colitis setting.
Su1531 Narrow Band Imaging Versus I-Scan for Real-Time Histologic Prediction of Diminutive Colonic Polyps: A Prospective Comparative Study Chang Kyun Lee1, Suck-Ho Lee2, Young Hwangbo3, Il Kwun Chung2, Tae Hoon Lee2, Sang-Heum Park2, Sun-Joo Kim2 1 Department of Internal Medicine, Kyung Hee University School of Medicine, Seoul, Republic of Korea; 2Department of Internal Medicine, Soon Chun Hyang University College of Medicine, Cheonan, Republic of Korea; 3Department of Preventive Medicine, Soon Chun Hyang University College of Medicine, Cheonan, Republic of Korea Background: Several recent studies have indicated that digital chromoendoscopy, such as narrow band imaging (NBI), Fujinon intelligent color enhancement, and I-Scan, can be useful for differentiating adenomatous from hyperplastic colonic polyps. However, there are no data demonstrating superiority of one system over another. Aim: The aim of this study was to compare the diagnostic efficacy of NBI with that of I-Scan for the real-time histologic prediction of diminutive colonic polyps (ⱕ 5 mm). Patients and Methods: A total of 142 consecutive patients undergoing screening or surveillance colonoscopy were prospectively enrolled from May to October 2010 (Clinical trial registration number: NCT1133041). Colonoscopy with real-time prediction of the polyp histology was performed by a single experienced endoscopist, with the use of commercially available high-definition endoscope systems equipped with digital chromoendoscopy technologies (Olympus NBI or Pentax I-Scan). All diminutive polyps detected during the procedure were carefully evaluated with highdefinition white light (HDWL), and thereafter with digital chromoendoscopy without optical and electronic magnification for the surface details. The polyp histology was predicted according to the polyp patterns that were reviewed and standardized before starting the study. Subsequently, the histology of all polyps was confirmed by evaluation of the biopsy samples or polypectomy specimens. All recorded endoscopic images were reviewed by a blind reader experienced in NBI and I-Scan technologies to evaluate inter-observer agreement of polyp differentiation. Results: A total of 156 polyps (80 adenomas and 76 hyperplastic polyps) from 70 patients were evaluated using the Olympus HDWL/NBI system (NBI group). There were 140 polyps (74 adenomas and 66 hyperplastic polyps) from 72 patients evaluated by the Pentax HDWL/I-Scan system (I-Scan group). There were no significant differences in the demographic data and basic polyp characteristics between the two groups. Both NBI and I-Scan had a significantly higher sensitivity and better accuracy than HDWL for prediction of adenomatous polyps (all P ⬍ 0.05). However, there were no significant differences between NBI and I-Scan for predicting adenomas (sensitivity, 88.8% vs 94.6%; specificity, 86.8% vs 86.4%; accuracy, 87.8% vs 90.7%, respectively; all P ⬍ 0.05). The level of intra-observer and inter-observer agreement was good (NBI, ⫽ 0.79 and 0.73, respectively; I-Scan, ⫽ 0.72 and 0.75, respectively). Conclusions: NBI and I-Scan have similar efficacy for real-time histologic prediction of diminutive colonic polyps. Both technologies are superior to HDWL colonoscopy for polyp differentiation. Limitations: The study did not have a “back-to-back” design and the study population was not randomized. Validation of the polyp classification system is needed.
www.giejournal.org
Su1532 Clinical Significance of Early Surveillance Colonoscopy After Endoscopic Removal of Early Colon Cancer Hyun Gun Kim, Jin-Oh Kim, Tae Hee Lee, Seong Ran Jeon, Ji Ho Ahn, Eun Jung Kang, Bong Min Ko, Joo Young Cho, Won Young Cho, Wan Jung Kim, Moon Sung Lee, Joon Seong Lee Institute for Digestive Research, Department of Gastroenterology and Hepatology, Soonchunhyang University, College of Medicine, Seoul, Republic of Korea Background and aims: The goal of surveillance colonoscopy after colon cancer resection is to detect recurrent cancer and metachronous lesions, and it is usually recommended 1 year after surgical resection. However, this is based on cases of surgically removed colon cancer, and limited data are available on the usefulness of surveillance colonoscopy after endoscopic cancer removal. We investigated the clinical significance and appropriate timing of surveillance colonoscopy after endoscopic cancer removal through an analysis of the outcome of first surveillance colonoscopy. Methods: We conducted a retrospective cohort study with a review of the clinical records of patients treated at a single tertiary medical center. From March 2004 to March 2009, 205 patients were diagnosed with early colon cancer after endoscopic mucosal resection or polypectomy. Of these, 139 patients who underwent surveillance colonoscopy within 1 year were enrolled. Patients who underwent surgery because of incomplete endoscopic cancer removal or aggressive pathological features and patients with inflammatory bowel disease, hereditary non-polyposis colon cancer syndrome, or familial adenomatous polyposis were excluded. All visible lesions, including cancer, were removed during endoscopic cancer removal at the first colonoscopy. Polyps newly detected at surveillance colonoscopy were analyzed in terms of their endoscopic appearance, shape, location, and histology, and we considered these lesions to be polyps missed at the first colonoscopy. We investigated the risk factors for missing an advanced lesion (defined as an advanced adenoma or cancer). Results: Overall, 774 lesions, including 159 colorectal cancers, were removed during the initial endoscopy for cancer removal. The mean interval to the first surveillance colonoscopy after cancer removal was 3.4 months (range 1-10). An additional 222 lesions were newly detected at surveillance colonoscopy. The overall miss rate of the initial endoscopy was 22.3%, and the miss rates for advanced adenoma and cancer (intramucosal cancer) were 11.4% and 3.6%, respectively. The miss rate was the highest for sessile polyps (28.3%) based on the endoscopic appearance, right colon polyps (29.6%) based on location, and polyps smaller than 1 cm based on size (p ⬍ 0.05). The total number of polyps seen during the initial colonoscopy was an associated risk factor for missing an advanced adenoma or cancer (odds ratio 1.176, 95% confidence interval 1.062-1.303, p ⫽ 0.002). Conclusions: Synchronous malignant or pre-malignant lesions can be missed during endoscopic cancer removal. The detection of a missed synchronous lesion highlights the clinical importance of the first surveillance colonoscopy after endoscopic cancer removal. The first surveillance colonoscopy may need to be performed earlier to reduce the incidence of missed synchronous lesions.
Su1533 Endoscopic Mucosal Resection of Colonic Polyps: Results From a Large Prospective Series Gaius R. Longcroft-Wheaton, Robert Mead, Moses Duku, Pradeep Bhandari Gastroenterology, Portsmouth Hospitals NHS trust, Portsmouth, United Kingdom Introduction: The traditional approach to the management of large colonic polyps has been surgery. Endoscopic mucosal resection (EMR) is an emerging technique for the removal of large colorectal lesions. Most of the published literature comes from Japan, with limited data regarding safety, efficacy and outcome in the west. We aim to assess the feasibility and safety of EMR in the colon in a western setting. Methods: A prospective review of patients undergoing EMR of colonic neoplasia ⬎2cm in size was performed. All patients were tertiary referrals from experienced consultants. The polyps were considered technically challenging due to size, difficult lesion access (peri-diverticular, periappendicular, touching the dentate line), or recurrences on previous EMR scars. They were referred to our service prior to surgical referral. Lesions were assessed using indigocarmine chromoendoscopy, and lesions with features suggestive of invasive malignancy were excluded. Completeness of resection was recorded by the endoscopist. Patients were followed up endoscopically where appropriate to assess for incomplete resection or recurrence. Results: 214 patients with 214 polyps underwent EMR. The mean size was 43mm (range 20-150). 180 were flat and 46 were on the right side of the colon. Primary reason for referral was the size in 91 cases, lesion access in 107 cases and a previous failed resection in 16 cases. Endoscopic clearance at first attempt was achieved in 92% of cases. Residual or recurrent disease was seen at the first endoscopic follow up in 17% cases requiring further endoscopic resection. Overall endoscopic cure has been achieved in 95% of patients. 3 patients went to surgery due to failed endoscopic resection. There were procedure related complications in 15/214 (7%) of cases. This consisted of delayed bleeding in 9 patients, immediate bleeding in 2 cases
Volume 73, No. 4S : 2011
GASTROINTESTINAL ENDOSCOPY
AB295
discrete lesions or flat mucosa). Seventy-five percent (33/44) had been on 5-ASA agents for a median of 9 years. Median total procedure time was 50mins. The median total examination time was identical with NBI and CE at 13mins, P⫽0.22. Overall 144 colonic lesions were detected, 21 of which were dysplastic. CE detected 131 lesions compared to 102 for NBI, P⬍0.015. There was no statistically significant difference in the number of dysplastic lesions detected by CE (20/21) and NBI (17/21). Fifty-five percent (79/144) of the lesions were 5mm or less in size. Most lesions were flat (Paris II: 53%, 76/144) or sessile (Paris Is: 36%, 52/144). The Kudo pattern was benign (I or II) in 84% (92/132) of lesions. Twenty-nine lesions (20%) had advanced Kudo patterns (IIIL, 26/144; IV, 3/144). The accuracy of the Kudo classification by NBI for neoplasia was 74% (95% CI: 0.68-0.80). In lesions occurring in areas of inflammation, Kudo accuracy for neoplasia was lower at 44% (95% CI: 0.35-0.51) compared to the rest of the lesions where accuracy was 75% (95% CI: 0.71-0.82). Summary: Chromoendoscopy identifies more discrete lesions during surveillance colonoscopy than narrow band imaging, and may be the preferred modality. Most lesions detected, however, are non-dysplastic. Most lesions are flat or sessile. The morphological Kudo classification appears to be less accurate at correlating with histology than in the non-colitis setting.
Su1531 Narrow Band Imaging Versus I-Scan for Real-Time Histologic Prediction of Diminutive Colonic Polyps: A Prospective Comparative Study Chang Kyun Lee1, Suck-Ho Lee2, Young Hwangbo3, Il Kwun Chung2, Tae Hoon Lee2, Sang-Heum Park2, Sun-Joo Kim2 1 Department of Internal Medicine, Kyung Hee University School of Medicine, Seoul, Republic of Korea; 2Department of Internal Medicine, Soon Chun Hyang University College of Medicine, Cheonan, Republic of Korea; 3Department of Preventive Medicine, Soon Chun Hyang University College of Medicine, Cheonan, Republic of Korea Background: Several recent studies have indicated that digital chromoendoscopy, such as narrow band imaging (NBI), Fujinon intelligent color enhancement, and I-Scan, can be useful for differentiating adenomatous from hyperplastic colonic polyps. However, there are no data demonstrating superiority of one system over another. Aim: The aim of this study was to compare the diagnostic efficacy of NBI with that of I-Scan for the real-time histologic prediction of diminutive colonic polyps (ⱕ 5 mm). Patients and Methods: A total of 142 consecutive patients undergoing screening or surveillance colonoscopy were prospectively enrolled from May to October 2010 (Clinical trial registration number: NCT1133041). Colonoscopy with real-time prediction of the polyp histology was performed by a single experienced endoscopist, with the use of commercially available high-definition endoscope systems equipped with digital chromoendoscopy technologies (Olympus NBI or Pentax I-Scan). All diminutive polyps detected during the procedure were carefully evaluated with highdefinition white light (HDWL), and thereafter with digital chromoendoscopy without optical and electronic magnification for the surface details. The polyp histology was predicted according to the polyp patterns that were reviewed and standardized before starting the study. Subsequently, the histology of all polyps was confirmed by evaluation of the biopsy samples or polypectomy specimens. All recorded endoscopic images were reviewed by a blind reader experienced in NBI and I-Scan technologies to evaluate inter-observer agreement of polyp differentiation. Results: A total of 156 polyps (80 adenomas and 76 hyperplastic polyps) from 70 patients were evaluated using the Olympus HDWL/NBI system (NBI group). There were 140 polyps (74 adenomas and 66 hyperplastic polyps) from 72 patients evaluated by the Pentax HDWL/I-Scan system (I-Scan group). There were no significant differences in the demographic data and basic polyp characteristics between the two groups. Both NBI and I-Scan had a significantly higher sensitivity and better accuracy than HDWL for prediction of adenomatous polyps (all P ⬍ 0.05). However, there were no significant differences between NBI and I-Scan for predicting adenomas (sensitivity, 88.8% vs 94.6%; specificity, 86.8% vs 86.4%; accuracy, 87.8% vs 90.7%, respectively; all P ⬍ 0.05). The level of intra-observer and inter-observer agreement was good (NBI, ⫽ 0.79 and 0.73, respectively; I-Scan, ⫽ 0.72 and 0.75, respectively). Conclusions: NBI and I-Scan have similar efficacy for real-time histologic prediction of diminutive colonic polyps. Both technologies are superior to HDWL colonoscopy for polyp differentiation. Limitations: The study did not have a “back-to-back” design and the study population was not randomized. Validation of the polyp classification system is needed.
www.giejournal.org
Su1532 Clinical Significance of Early Surveillance Colonoscopy After Endoscopic Removal of Early Colon Cancer Hyun Gun Kim, Jin-Oh Kim, Tae Hee Lee, Seong Ran Jeon, Ji Ho Ahn, Eun Jung Kang, Bong Min Ko, Joo Young Cho, Won Young Cho, Wan Jung Kim, Moon Sung Lee, Joon Seong Lee Institute for Digestive Research, Department of Gastroenterology and Hepatology, Soonchunhyang University, College of Medicine, Seoul, Republic of Korea Background and aims: The goal of surveillance colonoscopy after colon cancer resection is to detect recurrent cancer and metachronous lesions, and it is usually recommended 1 year after surgical resection. However, this is based on cases of surgically removed colon cancer, and limited data are available on the usefulness of surveillance colonoscopy after endoscopic cancer removal. We investigated the clinical significance and appropriate timing of surveillance colonoscopy after endoscopic cancer removal through an analysis of the outcome of first surveillance colonoscopy. Methods: We conducted a retrospective cohort study with a review of the clinical records of patients treated at a single tertiary medical center. From March 2004 to March 2009, 205 patients were diagnosed with early colon cancer after endoscopic mucosal resection or polypectomy. Of these, 139 patients who underwent surveillance colonoscopy within 1 year were enrolled. Patients who underwent surgery because of incomplete endoscopic cancer removal or aggressive pathological features and patients with inflammatory bowel disease, hereditary non-polyposis colon cancer syndrome, or familial adenomatous polyposis were excluded. All visible lesions, including cancer, were removed during endoscopic cancer removal at the first colonoscopy. Polyps newly detected at surveillance colonoscopy were analyzed in terms of their endoscopic appearance, shape, location, and histology, and we considered these lesions to be polyps missed at the first colonoscopy. We investigated the risk factors for missing an advanced lesion (defined as an advanced adenoma or cancer). Results: Overall, 774 lesions, including 159 colorectal cancers, were removed during the initial endoscopy for cancer removal. The mean interval to the first surveillance colonoscopy after cancer removal was 3.4 months (range 1-10). An additional 222 lesions were newly detected at surveillance colonoscopy. The overall miss rate of the initial endoscopy was 22.3%, and the miss rates for advanced adenoma and cancer (intramucosal cancer) were 11.4% and 3.6%, respectively. The miss rate was the highest for sessile polyps (28.3%) based on the endoscopic appearance, right colon polyps (29.6%) based on location, and polyps smaller than 1 cm based on size (p ⬍ 0.05). The total number of polyps seen during the initial colonoscopy was an associated risk factor for missing an advanced adenoma or cancer (odds ratio 1.176, 95% confidence interval 1.062-1.303, p ⫽ 0.002). Conclusions: Synchronous malignant or pre-malignant lesions can be missed during endoscopic cancer removal. The detection of a missed synchronous lesion highlights the clinical importance of the first surveillance colonoscopy after endoscopic cancer removal. The first surveillance colonoscopy may need to be performed earlier to reduce the incidence of missed synchronous lesions.
Su1533 Endoscopic Mucosal Resection of Colonic Polyps: Results From a Large Prospective Series Gaius R. Longcroft-Wheaton, Robert Mead, Moses Duku, Pradeep Bhandari Gastroenterology, Portsmouth Hospitals NHS trust, Portsmouth, United Kingdom Introduction: The traditional approach to the management of large colonic polyps has been surgery. Endoscopic mucosal resection (EMR) is an emerging technique for the removal of large colorectal lesions. Most of the published literature comes from Japan, with limited data regarding safety, efficacy and outcome in the west. We aim to assess the feasibility and safety of EMR in the colon in a western setting. Methods: A prospective review of patients undergoing EMR of colonic neoplasia ⬎2cm in size was performed. All patients were tertiary referrals from experienced consultants. The polyps were considered technically challenging due to size, difficult lesion access (peri-diverticular, periappendicular, touching the dentate line), or recurrences on previous EMR scars. They were referred to our service prior to surgical referral. Lesions were assessed using indigocarmine chromoendoscopy, and lesions with features suggestive of invasive malignancy were excluded. Completeness of resection was recorded by the endoscopist. Patients were followed up endoscopically where appropriate to assess for incomplete resection or recurrence. Results: 214 patients with 214 polyps underwent EMR. The mean size was 43mm (range 20-150). 180 were flat and 46 were on the right side of the colon. Primary reason for referral was the size in 91 cases, lesion access in 107 cases and a previous failed resection in 16 cases. Endoscopic clearance at first attempt was achieved in 92% of cases. Residual or recurrent disease was seen at the first endoscopic follow up in 17% cases requiring further endoscopic resection. Overall endoscopic cure has been achieved in 95% of patients. 3 patients went to surgery due to failed endoscopic resection. There were procedure related complications in 15/214 (7%) of cases. This consisted of delayed bleeding in 9 patients, immediate bleeding in 2 cases
Volume 73, No. 4S : 2011
GASTROINTESTINAL ENDOSCOPY
AB295