PTH-045a The Accuracy of Wavstat Version 4 Optical Biopsy Forceps in Characterising Colorectal Polyps Less 10 MM: A Prospective Blinded Study: Abstract PTH-045a Table 1

Abstract

IntroductionWavSTAT version 4 is an optical biopsy system designed for prediction of histology based on laser induced autofluorescence spectroscopy. The primary aim of this study was to demonstrate the accuracy of WavSTAT4 in characterising colorectal polyps <10 mm. The secondary aim was to compare the real time diagnostic performance of WavSTAT4 with NBI and a combination of endoscopic and WavSTAT assessments.Abstract PTH-045a Table 1Diagnostic performance of WavSTAT4, Endosocpic assessment and combined alogithmic assessment for characterisation of colorectal poylps < 10 mm and prediction of surveillance intervalsWavSTAT aloneWLE+NBI assessmentCombination of WavSTAT + endoscopic assessment (algorithmic approach)Sensitivity97.6%85.0%95.8%Specificity46.9%77.2%78%NPV96.8%91%98.5%PPV54.7%66%89.3%Surveillance interval (% of patients coded correctly)81.2%97%100%Surveillance interval (% of patients called earlier)18.8%3%0%MethodsAdult patients referred for lower gastrointestinal endoscopy were included in the study. Patients with inflammatory bowel disease or colorectal cancer were excluded. Polyps sized <10 mm were assessed in real time by high definition white light, NBI and WavSTAT4 optical biopsy forceps. Histopathological specimens were read separately by two expert GI pathologists blinded to the results of the NBI and WavSTAT assessments.Results156 polyps were found in 70 patients (Males-44, females-27, average age 65). After applying exclusion criteria a total of 126 polyps <10 mm were included in the analysis.Wavstat4 had a NPV of 96.8% but lacked specificity. Endoscopic assessment had a NPV of 91% and was more specific. Since the specificity of WavSTAT was poor mainly for hyperplastic recto-sigmoid polyps we evaluated an algorithmic approach where we classified the polyps according to the WavSTAT4 result when proximal to the recto-sigmoid junction. We classed them according to the endoscopic classification if WavSTAT4 predicted an adenomatous polyp in the recto-sigmoid area.This combined algorithmic approach met the PIVI thresholds and had a NPV of 95.8% and predicted 100% of surveillance intervals correctly.ConclusionWavSTAT version 4 has a high NPV for characterising colorectal polyps less than 10 mm in size but only predicts surveillance intervals correctly in 81.2% of patients. . An algorithmic approach combining Wavstat4 and endoscopic assessment had a high NPV with accurate prediction of surveillance intervals.Disclosure of InterestNone Declared