Analysis of small-bowel capsule endoscopy reading by using Quickview mode: training assistants for reading may...

The aim was to investigate the clinical utility of RAPID Access 6.5 Quickview software and to evaluate whether preview of the capsule endoscopy video by a trained nurse could detect significant lesions accurately compared with endoscopists.

As reading capsule endoscopy is time consuming, one possible cost-effective strategy could be the use of trained nonphysicians or newly available software to preread and identify potentially important capsule images.

The 100 capsule images of a variety of significant lesions from 87 patients were investigated. The minimum percentages for settings of sensitivity that could pick up the selected images and the detection rate for significant lesions by a well-trained nurse, two endoscopists with limited experience in reading, and one well-trained physician were examined.

The frequency of the selected lesions picked up by Quickview mode using percentages for sensitivity settings of 5%, 15%, 25%, and 35% were 61%, 74%, 93%, and 98%, respectively. The percentages for sensitivity significantly correlated (r=0.78, P<0.001) with the reading time. The detection rate by the nurse or the well-trained physician was significantly higher than that by the physician with limited capsule experience (87% and 84.1% vs. 62.7%; P<0.01). The clinical use of Quickview at 25% did not significantly improve the detection rate.

Quickview mode can reduce reading time but has an unacceptably miss rate for potentially important lesions. Use of a trained nonphysician assistant can reduce physician's time and improve diagnostic yield.