Investigating endoscopic features of sessile serrated adenomas/polyps by using narrow-band imaging with optical magnification

Background A sessile serrated adenoma/polyp (SSA/P) is a common type of colorectal polyp that possesses malignant potential. Although narrow-band imaging (NBI) can easily differentiate neoplastic lesions from hyperplastic polyps (HPs), SSA/Ps can be a challenge to distinguish from HPs. Objective To investigate specific endoscopic features of SSA/Ps by using NBI with optical magnification. Design Retrospective study. Setting Single high-volume referral center. Patients A total of 289 patients with histopathologically proven SSA/Ps or HPs obtained from colonoscopic polypectomy. Intervention Endoscopic images obtained by using NBI with optical magnification of 242 lesions (124 HPs, 118 SSA/Ps) removed between January 2010 and December 2012 were independently evaluated by 2 experienced endoscopists. Three external experienced endoscopists systematically validated the diagnostic accuracies by using 40 lesions (21 HPs and 19 SSA/Ps) removed between January and March 2013. Main Outcome Measurements Specific endoscopic features of SSA/Ps by using 5 potential characteristics: dilated and branching vessels (DBVs), irregular dark spots, a regular network pattern, a disorganized network pattern, and a dense pattern. Results Multivariate analysis demonstrated that DBV had a 2.3-fold odds ratio (95% confidence interval, 0.96-5.69) among SSA/Ps compared with HPs (sensitivity, 56%; specificity, 75%; accuracy, 65%). Interobserver and intraobserver agreement indicated almost perfect agreement for DBVs in both the evaluation and validation studies. When DBVs, proximal location, and tumor size (≥10 mm) were combined, the positive predictive value was 92% and the area under the curve was 0.783 in the receiver-operating characteristics by using the validation group. Limitations Retrospective study. Conclusions The current study suggests that a DBV is a potentially unique endoscopic feature of a colorectal SSA/P.