Article Text

PDF

Endoscopy II
PTU-229 The effect of frame rate and view mode on lesion detectability by novice and expert capsule endoscopists during reading
  1. M Nakamura,
  2. A Murino,
  3. A Fitzpatrick,
  4. Y Komeda,
  5. R La Nauze,
  6. S Green,
  7. C Fraser
  1. The Wolfson Unit for Endoscopy, St Mark's Hospital and Academic Institute, Imperial College, London, UK

Abstract

Introduction The RAPID 7 Access reading software (Given Imaging Ltd) allows the capsule endoscopist to adjust the frame rate of presented images (adjustable frame rate, AFR) and their view mode (VM1 - single view; VM2 - dual view; VM4 - quad view) during capsule endoscopy (CE) reading. The aim of this study was to establish the relationship between AFR, VM, lesion miss rate and reading time between non-expert (NEXs) and expert (EXs) capsule endoscopists.

Methods One short video clip containing 60 positive images of angioectasias was selected from our CE database. The clip was read by 3 EXs and 3 NEXs using nine different combinations of VM and AFR (1, 2 and 4 VMs × 10, 15 and 25 fps) presented in randomised order. Readers were asked to count each positive image of an angioectasia using a manual counter, without interrupting the video clip.

Results The reading times at 10, 15 and 25 fps were 54, 34 and 20 s, respectively for any VM. Considering 10 fps as the gold standard, an AFR of 15 and 25 fps resulted in a reduction in reading time of 37% and 63% respectively. The number of positive images detected using 10, 15 and 25 fps (all VM's combined) were 45, 31 and 22 respectively. The mean number of detected positive images (MPI) using 10 fps was significantly higher than an AFR of 15 and 25 fps (p=0.04, 0.01). For VM1, VM2 and VM4, the MPI was 24, 36 and 38 respectively. The MPI using VM2 and VM4 was significantly higher than for VM1 (p=0.01, 0.003). VM4 × 10 fps had highest MPI (51) while VM1 × 25 fps had the lowest MPI (14). MPIs of NEX and EX (all VM's combined) were 34 and 32 and were not significantly different.

Conclusion While a higher AFR results in a reduction in reading time, lesion detectability is reduced and miss rates increase. Higher MPIs are associated with lower AFRs and an increase in VM. In this study the optimal combination for lesion detectability was VM4 × 10 fps. NEXs and EXs performed similarly for the detection of angioectasias.

Competing interests None declared.

Statistics from Altmetric.com

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.