TBA (21W122)

The true false negative rate of colon capsule endoscopy (CCE) is low.

Author(s)

Semenov S*, Ismail MS*, O’Donnell S, O’Connor A, Breslin N, Ryan B, McNamara D*

Department(s)/Institutions

Department of Gastroenterology, Tallaght University Hospital Trinity Academic Gastroenterology Group, Trinity College Dublin*

Introduction

CCE is an established diagnostic tool for colonic pathology. There is a lack of clinical data on true capsule false negative (FN) rates.

Aims/Background

To assess the causes of missed pathology in a CCE cohort.

Method

CCEs with at least one colonoscopy within 18 months were identified from a 9-year database. Missed pathology on index capsules was identified by comparing with colonoscopy. Indication, bowel preparation, missed lesion/pathology characteristics were collated. These studies were re-read by experts unblinded to CCE and colonoscopy findings. On re-reading, newly identified lesions/pathology, verified by an expert panel, were categorised as reader error, the remainder were considered true FN events.

Results

Of 532 CCEs, 210 (39%) had a comparative colonoscopy (mean interval 4 months) and 49 (23%) had missed pathology; 30/49 (61%) reaching the colonic section with missed pathology. Of 30 discrepant studies, 24 (80%) had adequate preparation. Indications included 14 (47%) polyp surveillance, 12 (40%) GI symptoms, 3 (10%) IBD and 1 (3%) screening. Missed pathology included diminutive polyps 18 (60%), polyps ≥6mm 8 (27%), inflammation 4 (13%); they were evenly distributed (14 right and 16 left colon) and 18/26 (69%) were adenomas. No cancers were missed. Reader error accounted for 23/30 (77%) cases, while 7/30 (23%) were true FN events. Missed pathology was due to true capsule error 7/210 (3.3%), reader error 23/210 (11%), incomplete studies 19/210 (9%). Reader error was more likely than capsule error, OR 3.2 (p=0.018, 95%CI 1.22-8.80).

Conclusions

Our study suggests a low true CCE FN rate. Reader error accounts for most missed pathology.