ORIGINAL ARTICLE Comparison of high-resolution magnification narrow-band imaging and white-light endoscopy in the prediction of histology in Barrett’s oesophagus RAJVINDER SINGH 1 , HARIS KARAGEORGIOU 1 , VICTORIA OWEN 2 , KLARA GARSED 1 , PAUL J. FORTUN 1 , EDWARD FOGDEN 1 , VENKATARAMAN SUBRAMANIAM 1 , ANTHONY SHONDE 1 , PHILIP KAYE 3 , CHRISTOPHER J. HAWKEY 1 & KRISH RAGUNATH 1 1 Wolfson Digestive Diseases Centre, Nottingham University Hospitals NHS Trust, Nottingham, 2 Trent Research Development Unit, University of Nottingham, and 3 Department of Histopathology, Nottingham University Hospitals NHS Trust, Nottingham, UK Abstract Objective. To evaluate whether there is any appreciable difference in imaging characteristics between high-resolution magnification white-light endoscopy (WLE-Z) and narrow-band imaging (NBI-Z) in Barrett’s oesophagus (BE) and if this translates into superior prediction of histology. Material and methods. This was a prospective single-centre study involving 21 patients (75 areas, corresponding NBI-Z and WLE-Z images) with BE. Mucosal patterns (pit pattern and microvascular morphology) were evaluated for their image quality on a visual analogue scale (VAS) of 110 by five expert endoscopists. The endoscopists then predicted mucosal morphology based on four subtypes which can be visualized in BE. Type A: round pits, regular microvasculature; type B: villous/ridge pits, regular microvasculature; type C: absent pits, regular microvasculature; type D: distorted pits, irregular microvasculature. The sensitivity (Sn), specificity (Sp), positive predictive value (PPV), negative predictive value (NPV) and accuracy (Acc) were then compared with the final histopathological analysis and the interobserver variability calculated. Results. The overall pit and microvasculature quality was significantly higher for NBI-Z, pit: NBI-Z 6, WLE-Z 4.5, p B0.001; microvasculature: NBI-Z 7.3, WLE- Z 4.9, p B0.001. This translated into a superior prediction of histology (Sn: NBI-Z: 88.9, WLE-Z: 71.9, p B0.001). For the prediction of dysplasia, NBI-Z was superior to WLE-Z ( x 2 10.3, p B0.05). The overall kappa agreement among the five endoscopists for NBI-Z and WLE-Z, respectively, was 0.59 and 0.31 ( p B0.001). Conclusions. NBI-Z is superior to WLE-Z in the prediction of histology in BE, with good reproducibility. This novel imaging modality could be an important tool for surveillance of patients with BE. Key Words: Barrett’s oesophagus, high resolution, magnification, narrow-band imaging, white-light endoscopy Introduction Random 4-quadrant biopsies are presently the stan- dard surveillance method used for detection of dysplasia in Barrett’s oesophagus (BE). It is commonly known that apart from being tedious, uncomfortable and time-consuming, random biop- sies are associated with a sizeable sampling error. While investigating the current practice of surveil- lance strategy for patients with BE in the UK, Mandal et al. showed that only 41% of British gastroenterologists followed the random quadrantic biopsy protocol [1]. Recently, numerous novel endoscopic imaging modalities have been introduced in the hope that ’’optical biopsies’’ can be taken, hence increasing the yield of detecting dysplasia in BE, making surveillance less cumbersome and ultimately enabling the prolongation of surveillance intervals without increasing the incidence of interval Correspondence: Rajvinder Singh, The Lyell McEwin Hospital, Haydown Road, Elizabeth Vale, SA 5112 Australia. Tel: 61 8 81829909. E-mail: rajvinder.singh@health.sa.gov.au Scandinavian Journal of Gastroenterology, 2009; 44: 8592 (Received 26 November 2007; accepted 3 August 2008) ISSN 0036-5521 print/ISSN 1502-7708 online # 2009 Informa UK Ltd. DOI: 10.1080/00365520802400818