Polyps

Abstracts

Morphology Pedunculated Sessile Flat

Removal technique Snare polypectomy Lift and snare Forceps

SPs (%)

ADs (%)

P-value

RR

95% CI

29 (2.0) 1262 (80.6) 274 (17.4) 505 (23.9) 693 (32.8) 918 (43.3)

214 (14.2) 1210 (80.5) 79 (5.3) 680 (40.9) 351 (21.2) 630 (37.9)

!0.001 0.93 !0.001

0.11 1.00 3.33

0.07-0.17 0.97-1.04 2.62-4.23

!0.001 !0.001 0.001

0.13 1.55 1.14

0.09-0.19 1.39-1.73 1.06-1.24

Table 2. Risk analysis for polypectomy complications per polyp Histology (SP vs. AD) Colon location (proximal vs. distal) Size (R10mm vs. !10mm) Morphology (non- vs. pedunculated) Removal technique (snare vs. forceps) SPS vs. non-SPS patients

RR

95% CI

P-value

0.41 1.50 16.40 0.60 0.93 1.05

0.10-1.71 0.36-6.27 3.93-68.38 0.07-4.90 0.21-4.14 0.26-4.20

0.22 0.58 !0.001 0.64 0.92 0.94

Tu1416 Sessile Serrated Polyp Detection Rate in a Large Community Practice As a Potential Quality Indicator for Colonscopy Carson Keck*1, Mudit Chowdhary1, Matthew N. Varn1, Minh Hang1, Ali Keshavarzian2, Shahriar Sedghi1 1 Mercer University School of Medicine, Macon, GA; 2Rush University Medical Center, Chicago, IL Background: Adenoma detection rate is currently being used as one of the quality measures in colonoscopy performance. However, this criterion may be flawed because a high initial detection rate may subsequently lead to a lower detection rate during follow-up surveillance. We propose that sessile serrated polyp detection rate (SSPDR) may be a better quality measure as these polyps are harder to detect, have a higher potential risk for malignancy, and may recur more rapidly. Furthermore, for the same reason, SSPDR may be better in comparing detection sensitivity between colonoscopy and emerging technologies such as virtual colonoscopy and capsule technology. Surprisingly, there is scant information regarding SSPDR and these studies show a variable detection rate, which may be due to inconsistent pathological diagnostic criteria. Objective: The aim of our study was to compare the SSPDR with the adenoma detection rate (ADR) in approximately 13,000 patients at a large outpatient center with multiple physicians over a 2-year period. Additionally, the colon withdrawal time in minutes was compared to SSPDR and ADR. Methods: This was a comparative, retrospective study in which 277 patients had an SSP detected during colonoscopy over a 2-year span. The ADR and the SSPDR were calculated for each physician and for the practice as a whole. A Pearson’s correlation coefficient was then calculated between the SSPDR and the ADR. The average withdrawal time for each doctor was calculated and a Pearson’s correlation coefficient was computed between the withdrawal times and both the sessile serrated polyp and adenoma detection rates. Results: The mean adenoma detection rate was 25.64 and the mean sessile serrated polyp rate was 1.95. The mean withdrawal time was recorded as 9.73 minutes (See Table 1). The Pearson’s correlation coefficient between the ADR and SSPDR was RZ 0.756. Analysis of the SSPDR and withdrawal time revealed an R value of -0.141. An R value of 0.024 was calculated reflecting the correlation between ADR and withdrawal time. Conclusion: Statistical analysis revealed a positive correlation between ADP and SSPDR. A slight negative correlation was identified between the SSR detection rate and withdrawal time. This may reflect overall physician experience rather than withdrawal time alone. The low variance for SSPDR in our study may signify that detection of sessile serrated polyps may prove to be a more consistent quality measure than ADR. However, we recommend further studies to classify SSPDR as a quality measure as opposed to or in addition to ADR. If SSP detection is a more accurate quality measure, further studies would be needed to assess this in follow up surveillance colonoscopies. It would also be beneficial to compare sensitivity of SSPDR via colonoscopy and virtual colonoscopy and other emerging technologies. Table 1

Mean Variance Standard Deviation Standard Error

SSP Detection Rate

Adenoma Detection Rate

Withdrawal Time (minutes)

1.95 0.33 0.57 0.21

25.64 7.15 2.67 1.02

9.73 2.70 1.64 0.62

AB424 GASTROINTESTINAL ENDOSCOPY Volume 79, No. 5S : 2014

Tu1417 Prevalence of Serrated Adenomas: Experience in a Large Volume Centre in Argentina Eugenia Del Cantare*1, Ricardo F. Figueredo1, Sandra Canseco1, Mauricio Fuster1, Patricio Sheridan1, Jorge E. Bosch1, Alejandra Avagnina3, Boris Elsner4, Santiago De Elizalde2, Carolina Bolino1, Luis E. Caro1, Cecilio L. Cerisoli1 1 Gedyt, Buenos Aires, Argentina; 2Laboratorio de Anatomia Patologica, Buenos Aires, Argentina; 3Centro de Patologia Dr. Elsner, Buenos Aires, Argentina; 4CEMIC, Buenos Aires, Argentina Introduction: Serrated adenomas (SA) evolve to Colorectal Cancer (CRC) in 15-20% through the serrated pathway. The prevalence is variable (0.8 to 13%) because some lesions are unseen even in case of experienced endoscopists. Objectives: 1. To estimate the prevalence of SA. 2. To describe their endoscopic and histological characteristics. Materials and methods: AdultsR 18 years who performed videocolonoscopy (VCC) were included consecutively. Anticoagulation, incomplete studies, insufficient colon cleaning (Boston scale !6), high risk for CRC except for Inflammatory Bowel Disease were exclusion criteria. The study was conducted in an outpatient gastroenterology clinic between November, 2012 and April, 2013. Design: descriptive, prospective and cross sectional study. The VCC were performed under sedation with Olympus equipment and by trained operators. Colon cleansing was done with polyethylene glycol or phosphates, with / without bisacodyl. The lesions were resected according to current practice. Histological evaluation was performed according to the World Health Organization criteria including Sessile Serrated Adenomas (P / SSA), the P / SSA with dysplasia and Traditional Serrated Adenomas (TSA). Age, gender and risk factors for CRC were evaluated as confounding variables. Ethics: The protocol was approved by the local IRB. Statistical analysis: VCCstat 2.0. package, 95%CI; Chi square test. Results: We reviewed 3052 VCC; 316 were excluded and 2736 were analyzed. 58% (1584/2736) were female, mean age: 56  12.6 years (range 20-93). According to CRC risk, 4 groups were established: No risk: 18.45%, average: 58.9%, moderate: 21.8%, and high: 0.85%. 73% had no lesions. 100 polyps were recorded in 75 patients. No differences were observed between R50 and !50 years and between genders (p Z ns). A higher prevalence of SA was observed in patients with moderate risk (p Z 0.0048) and there was a tendency that these lesions were more prevalent at average risk group (p Z 0.0684) 1. The prevalence of SA was 2.7% (95%CI 2.2-3.4). 2. Most prevalent endoscopic features are detailed in the table. P/SSA were prevalent in 87% (95%CI 78-92) followed by P/SSA with dysplasia 13% (95%CI 7-21); all these latter lesions had Low Grade Dysplasia. Conclusions: In this sample, the prevalence of SA was low and SSA without dysplasia were predominant. Most of these lesions, regardless the dysplasia, were similar in size and morphology but differed in the location; P/SSA were more prevalent in recto sigmoid, followed by proximal colon and transverse, and P/SSA with dysplasia were more prevalent in transverse followed by proximal colon and rectosigmoid. This data should warn endoscopists to emphasize the importance of colonic cleansing and rigorous evaluation of right and transverse colon.

Predominant endoscopic and histologic features of SA Histological and endoscopic features n Size (mm) Morphology Location

P/SSA with dysplasia (n:13)

P/SSA (n:87) % 95CI

n

6-9 Is Transverse Recto sigmoid

50 50 32

% 95CI 57,5 (46-67) 57,5 (46-67) 36,8 (27-48)

9 8 6 -

69,2 (38-90) 61,5 (31-86) 46,2 (19-74) -

Tu1418 Risk Factors for Dysplasia in Sessile Serrated Adenomas / Polyps Leonardo C. Duraes*, David Liska, Matthew F. Kalady, James M. Church Colorectal Surgery, Cleveland Clinic Foundation, Cleveland, OH Purpose: Serrated lesions of the colorectum are the precursors of about onethird of colorectal cancers. The World Health Organization (WHO) classification of serrated colorectal polyps includes hyperplastic polyps, sessile serrated adenomas / polyps (SSA/P) with or without dysplasia, and traditional serrated adenomas (TSA). Among these lesions, SSA/P with dysplasia is the main precursor of carcinoma. The development of dysplasia in SSA/P has been linked to promoter methylation of MLH and microsatellite instability. Unfortunately, little is known about clinical factors that may predict dysplasia in SSA/P, and thus identifying patients at risk is difficult. The aim of this study was to identify factors that may predict dysplasia in SSA/P. Methods: A single colonoscopist database was queried to identify patients with

www.giejournal.org

Abstracts

histologically confirmed SSA/P, from 2003 to 2013. Since polyp size is a wellknown predictor of dysplasia in polyps, patients with SSA/Ps with dysplasia were matched against patients with SSA/Ps without dysplasia by polyp size (1:3). No control patient was used more than once. Patients with Serrated Polyposis Syndrome as defined by WHO criteria were excluded from analysis. Several factors were analyzed (see table) to determine association with dysplasia. Univariate statistical analysis was performed, and p!0.05 was considered significant. Results: 474 patients with SSA/Ps were identified, 448 patients with SSA/Ps without dysplasia (94.5%), and 26 with dysplasia (5.5%). In this initial group of patients, there was a significant difference in the size of SSA/P with dysplasia (16.6 mm) compared to SSA/P without dysplasia (12.2 mm) (pZ0.023). The 26 patients with SSA/P with dysplasia were then matched by SSA/P size against 78 patients who had SSA/P without dysplasia. The overall mean age was 63.4 years (63.1 without dysplasia, 64.3 with dysplasia). The overall mean polyp size was 18 mm  10.9 mm and the median was 20 mm (range 3-50 mm). Results are summarized in the table. Male gender and cumulative number of polyps were associated with dysplasia (pZ0.05, pZ0.031, respectively). No statistical difference was observed in age, polyp site, family history of colorectal cancer, indication for colonoscopy, or number of synchronous adenomas. There was no difference in the number of colonoscopies performed and in cancer incidence between the groups. Conclusion: SSA/P size is associated with an increased risk of dysplasia, and men with more colorectal polyps are at risk of developing dysplasia in sessile serrated adenomas / polyps. Colonoscopy in this group should be particularly uncompromising.

Patient characteristics by presence of dysplasia SSA/P without dysplasia

SSA/P with dysplasia

p Value

Initial patients SSA/P size (mm)

448 (94.5%) 16.6 (8.6)

26 (5.5%) 12.2 (8.7)

0.023

Matched patients SSA/P size (mm) Age Gender Female Male Location of polyps Right Colon Left Colon Rectum Indication for colonoscopy Screening Family History Cancer/ Polyps Past History Cancer/ Polyps Total number of polyps per patient Total number of SSA/P Total number of adenomas Total number of hyperplastic polyps Total number of Scopes Cancer

78 18.1 (10.9) 63.1

26 18.1 (11.1) 64.3

53 (67.9%) 24 (32.1%)

9 (34.6%) 17 (65.4%)

66 (84.6%) 11 (14.1%) 1 (1.3%)

22 (88.0%) 3 (12.0%) 0

0.989 0.680 0.005

1.000

0.297 17 (30.9%) 4 (7.3%)

2 (12.5%) 2 (12.5%)

34 (61.8%)

12 (75%)

5.3 (6.1)

9.2 (7.6)

0.031

2.0 (2.0) 1.9 (4.0)

3.2 ( 3.3) 3.0 (4.8)

0.152 0.325

1.1 (1.4)

1.8 (1.8)

0.066

2.1 (1.1) 1 (1.2%)

2.8 (1.7) 2 (7.7%)

0.066 0.274

Tu1419 Endoscopic Features and Management Outcomes of Colonic Sessile Serrated Polyps Murad Abu Rajab*, Joshua B. Max, Laith H. Jamil, Kapil Gupta, Simon K. Lo Pancreatico-biliary, Cedars Sinai Medical Center, Los Angeles, CA Introduction: Colonic sessile serrated lesions, first described in 1996, are typically flat lesions that require endoscopic mucosal resection (EMR). It is believed that up to 30% of colon cancers originate from sessile serrated polyps. Yet there are few published reports characterizing these lesions. We report our experience on the endoscopic features and management outcomes on these lesions. Methods: retrospective review of all colonoscopies referred for EMR in our institution from 9/2006 to 11/2013. All data were recorded prospectively in our procedure database. 57 patients underwent colonoscopy with the intent mucosal resection. One patient had very poor colonic prep-

www.giejournal.org

aration and was excluded from the study. Collected data included patients, age, sex, polyp location, size,pathology endoscopicappearance, resectiontechnique and residual neoplasia. Patients werefollowing with surveillance colonoscopy to ensure complete polyp removal. Results: 56 patients, with 82 polyps, were identified. 59% were women, mean age 65  12 SD years. Polyps were located in the right colon (nZ64; 77%), transverse colon (nZ13; 16%), left colon (nZ5; 6%), near ileo-colonic anastomosis (nZ1; 1%) and none in the rectum. The mean polyp size (defined as the longest dimension) was 1.9 cm (0.7-4.5 cm,). Polyp appearance included flat (nZ 64; 78%), slightly raised (nZ 11; 13.5%) and polypoid (nZ7; 8.5%). EMR technique was cap-assisted (CAM) in 52 polyps (63%) and Freehand (FHM) using spiral snare in 30 polyps (36%). 8 (10%) of the polyps not adequately resected using the freehand technique were switched to CAM to ensure complete removal. Complete polyp removal was achieved in 76 polyps (93%) in the initial session. CAM was superior to FHM in complete polyp removal (96.6% vs 73%. Fisher’s exact test, (pZ0.002). Histopathology showed sessile serrated polyp (SSP) 75(91%), SSP with adenoma in 4(5%), SSP with low grade dysplasia in 1 (1%) and SSP with intramucosal carcinoma in 2(2.5%). Follow up colonoscopy was done on 40 patients (71%) (29 CAM, 8 FHM and 3CMA+FHM treated patients) with a mean follow up of 14 months. Residual tissue was found in 2 CAM-treated patients and they were removed. Complications were post-polypectomy bleed in 2/56 (3.6%), one stopped spontaneously and one required endoscopic intervention. 2(3.6%) patients had abdominal pain that required hospitalization and one ( 2%) patient had perforation that was managed with endoscopic clips and no surgical intervention was done. Conclusions: The vast majority (77%) of SSPs are located in the right colon, and most (78%) are flat without any raised edge. It is rare (2.5%) that they contain malignant components when discovered. CAM is superior to FHM in achieving complete removal of SSP, without encountering serious complications.

Tu1420 Risk of Metachronous Advanced Neoplastic Lesions in Patients With Sessile Serrated Adenomas Undergoing Surveillance Colonoscopy Lisandro Pereyra*1, Estanislao J. GóMez1, Rafael Zamora1, Carolina Fischer1, Guillermo N. Panigadi1, Raquel GonzáLez1, Maximiliano BUN2, Cristina I. Vucko Anriquez1, Paula Galletto1, José M. Mella1, Pablo Luna1, Silvia C. Pedreira1, Daniel G. Cimmino1, Luis a. Boerr1 1 Gastroenterology and Endoscopy Unit, Hospital Alemán, Buenos Aires, Argentina; 2Colorectal Surgery Section, Hospital Alemán, Buenos Aires, Argentina Introduction: Although sessile serrated adenomas (SSAs) may represent a separate and important pathway for colorectal cancer. The risk of methacronous colonic neoplastic lesions, in patients with SSA undergoing surveillance colonoscopy, is not well quantified. Objective: To compare the risk of metachronous advanced neoplastic lesions during colonoscopy surveillance in patients with SSA, high and low risk conventional adenomas, and negative colonoscopy. Methods: This was a single-site retrospective study of patients with index SSA, low risk adenomas (LRA), high risk adenomas (HRA); and negative index colonoscopy (NIC) (from January 2007 to December 2008), who underwent colonoscopy surveillance for a period of at least 3 years. SSA diagnosis was performed by two blinded gastrointestinal pathologist according to Snover criteria. HRA were defined as an advanced neoplastic lesion (ANL) (O75% villous histology, high grade dysplasia or size O1 cm) or R 3 non-advanced neoplastic lesions. Metachronous ANL was considered when occurring 12 month after index colonoscopy. Continuous variables were compared using one-way analysis of variance (ANOVA) and Kruskal Wallis. KapplanMeier curves and logrank test were used to evaluate time to first ANL during colonoscopy surveillance. Results: Among 639 included patients, index colonoscopy finding were: 75 SSA, 140 LRA, 87 HRA and 337 NIC. Patients with SSA and NIC were younger than patients with LRA and HRA: mean age (SD) 56 (10), 59 (8), 64 (11) and 65 (9) (p !0.01). The mean colonoscopy follow-up was longer in patients with NIC and LRA than SSA and HRA, months (SD): 63 (10), 56 (17), 47 (14), and 52 (17). The number of surveillance colonoscopies per patient was lower in patients with NIC (p!0.001). There were no difference between groups regarding to: gender, number of surveillance colonoscopies per patient and satisfactory bowel preparation on index colonoscopy (pO0.4). The prevalence of metachronous ANL in patients with SSA, LRA, HRA and NIC was: 12%, 8.5%, 22% and 1.5 respectively (logrank test!0.01) (Figure 1). Time to first metachronous ANL during colonoscopy surveillance in patients with SSA, LRA, HRA, and NIC was: months (SD) 47 (13), 55 (17), 50 (14) and 63 (10) (logrank test !0.01). Presence of synchronic conventional adenoma in patients with index SSA was associated with the highest prevalence of metachronic ANL during surveillance (34.6%) (logrank test !0.01). None of the patients with index SSA without synchronic adenoma presented a metachronous ANL during surveillance (Figure 2). Conclusion: Patients with SSA have an increased risk of presenting metachronous ANL during colonoscopy surveillance. This risk seems to be higher than with LRA but lower than with HRA. The SSA risk of

Volume 79, No. 5S : 2014 GASTROINTESTINAL ENDOSCOPY AB425