SMALL POLYPS AT COLONOSCOPY AND THE NICE CLASSIFICATION: LIKELY CAUSES OF OPTICAL DIAGNOSIS ERROR

IntroductionThere is increasing interest in optical diagnosis (OD) for small colonic polyps with a ‘resect and discard’ strategy. During OD the endoscopist assigns the polyp diagnosis as high or low confidence using the NICE or equivalent classification. Our aim is to assess factors that result in an incorrect OD where a high confidence optical diagnosis is made by an experienced endoscopist.MethodsIn the early phase of a prospective feasibility study of optical diagnosis (DISCARD3), eight bowel cancer screening colonoscopists optically diagnosed 639 diminutive polyps during Feb-Nov 2020. Each polyp diagnosis was evaluated by the colonoscopist as high or low confidence using the NICE criteria. All retrieved polyps were sent for histopathology. In 112 polyps the high confidence optical diagnosis did not match the final histology result. A root cause analysis was performed as part of the ongoing training and implementation of optical diagnosis. Each of these polyp photos were assessed by a blinded endoscopist who scored photo quality with 1 point for each of the following criteria (score range 0-3): clean surface; complete view; and correct focal distance. A score of 3 was considered adequate to apply the NICE criteria. In addition, the likely cause of the optical diagnosis error was categorised as: 1. NICE mismatch - NICE criteria could not be fully applied to the polyp (eg mixture of NICE type 1 and type 2 features). NICE not applied – NICE criteria could be applied but was done incorrectly. Inadequate photo for NICE – where photo quality was inadequate for NICE criteria to be applied. Specimen processing error – where there is high confidence that the optical diagnosis was correct with suspected specimen processing/retrieval error.ResultsOf 112 polyps, 91% (102/112) had clean surface, 60% (67/112) had complete view, and 27% (30/112) had correct focal distance. The photo quality scores were inadequate in 93/112 cases (score 0: 5%, 6/112; 1: 29%, 32/112; or 2: 49%, 55/112) and adequate in 19/112 cases (score 3: 17%, 19/112). The assumed cause of the optical diagnosis error was NICE mismatch in 31/112 (27.7%) cases, NICE not applied in 19/112 (17.0%), inadequate photo for NICE in 59/112 (52.7%) and histology processing error in 3/112 (2.7%).ConclusionsAdequacy of polyp photo is critical to quality assurance of OD. In this study a significant proportion of OD errors had corresponding inadequate photodocumentation with incorrect focal distance the main photographic issue. This may affect application of the NICE criteria. In cases where photos were adequate, understandable error occurred where lesions had a mixture of NICE 1 and 2 features.