Narrow-Band Imaging Improves the Detection of Polyps in Patients with Hyperplastic Polyposis Syndrome: A Prospective Randomized Study Comparing Narrow-Band Imaging vs High-Resolution White Light Imaging

Abstract

Introduction: Hyperplastic polyposis syndrome (HPS) is characterized by the presence of multiple colorectal hyperplastic polyps (HPs). Besides HPs, macroscopically similar sessile serrated adenomas (SSAs) and conventional adenomas are common findings in HPS and are believed to be the pre-malignant polyps. Detection and removal of these high-risk polyps may be an important step in preventing colorectal cancer development in HPS patients. However, polyps in HPS can be difficult to detect due to their unremarkable color and flat shape. Narrow-band imaging (NBI) may improve the detection and differentiation of these polyps. Aims and Methods: To prospectively compare the value of NBI and high-resolution white light endoscopy (WLE) for the detection of polyps in HPS and to assess the value of NBI for the differentiation of these polyps. Consecutive HPS patients undergoing surveillance colonoscopies were included in this study. Each colonic segment was inspected twice, once with WLE and once with NBI, in random order by one experienced endoscopist. Of all detected polyps the size, shape and location was assessed. Kudo pit-pattern analysis was performed in all polyps. Results: In this study, 22 patients were included, in whom a total of 182 polyps were detected: 116 HPs, 42 SSAs and 24 adenomas. The median size of polyps was 5mm (range: 2-20mm). Macroscopically, polyps were classified as flat (60%), sessile (38%) and pedunculated (2%). Proximal location was seen in 54% of HPs, 81% of SSAs and 92% of adenomas. Among patients assigned to inspection with WLE first (n=11), 70 polyps were detected. During subsequent inspection with NBI, 40 additional polyps were visualized. Among patients examined with NBI first (n=11), 65 polyps were detected. Subsequent inspection with WLE rendered 7 additional polyps. The sensitivities of WLE and NBI for overall polyp detection were 64% and 90% respectively (p<0.001). For flat polyps these were 51% and 87% (p<0.001) and for sessile/pedunculated polyps 81% and 96% (ns). The sensitivities of WLE and NBI for HPs was 67% and 90% respectively (p=0.017). For SSAs this was 38% and 86% (p=0.003) and for adenomas 70% and 100% (ns). On NBI, Kudo pit pattern type I-II was seen in 83 (76%) HPs versus 27 (73%) SSAs (ns). Only 6 (25%) adenomas had Kudo type I-II (p<0.001). The accuracy of NBI for discriminating SSAs from HPs was 63% and for differentiating adenomas from HPs this was 75%. Conclusion: NBI significantly improves the detection rate of high-risk SSAs and predominant flat polyps in HPS. Therefore, NBI should be considered as standard practice of care for the endoscopic surveillance of HPS patients.