University of Maryland School of Medicine, Baltimore, Maryland 21201, USA.
Am J Gastroenterol. 2012 Apr;107(4):554-60. doi: 10.1038/ajg.2011.461. Epub 2012 Jan 10.
Limited data guide capsule endoscopists on how to view the many images collected in each capsule. The objective of this study was to compare the detection rates of clinically significant findings in different capsule endoscopy reading modes and speeds.
Seventeen capsule endoscopists with experience from 23 to > 1,000 total capsule procedures read 24 clips, 18 of which were abnormal. Clips were read in two different reading modes utilizing two speeds, including SingleView at 15 at frames per second (f.p.s.), SingleView 25 f.p.s., QuadView 20 f.p.s., and QuadView 30 f.p.s. The main outcome measurements were pathology detection rates correlated with reading mode, lesion type, reader experience, and timing order.
SingleView 15, QuadView 20, and QuadView 30 had no significant difference in overall detection rate (45, 47, and 43%, respectively). SingleView 25 had a 26% detection rate, which was significantly lower than SingleView 15 (P = 0.04) and QuadView 20 (P = 0.002). The detection rates of angioectasias, ulcers/erosions, masses/polyps, and blood were 69, 38, 46, and 17%, respectively. Reader experience and timing of interpretation did not significantly impact detection rate.
Pathology was present on a few frames. Limited modes and speeds were assessed. Lesion types were not confirmed with surgical or deep enteroscopic methods. A relatively small number of readers provided interpretations.
Overall, the detection rates in this study are lower than previously reported and not influenced by increasing experience. Detection rates are significantly higher when reading in SingleView 15 and QuadView 20 compared with reading in SingleView 25. Increasing viewing speed from QuadView 20 to QuadView 30 appears to have no significant effect on detection. Quality control measures to compare and improve lesion detection rates need further study.
有限的数据指导胶囊内镜医生如何查看每个胶囊收集的大量图像。本研究的目的是比较不同胶囊内镜阅读模式和速度下对临床有意义的发现的检出率。
17 名经验丰富的胶囊内镜医生（总胶囊操作经验从 23 例到＞1000 例）阅读了 24 个视频片段，其中 18 个为异常片段。使用两种速度的两种阅读模式阅读视频片段，包括每秒 15 帧的单视图（SingleView）、每秒 25 帧的单视图（SingleView 25）、每秒 20 帧的四视图（QuadView）和每秒 30 帧的四视图（QuadView 30）。主要观察指标为与阅读模式、病变类型、医生经验和时间顺序相关的病理检出率。
SingleView 15、QuadView 20 和 QuadView 30 的总体检出率无显著差异（分别为 45%、47%和 43%）。SingleView 25 的检出率为 26%，显著低于 SingleView 15（P=0.04）和 QuadView 20（P=0.002）。血管扩张、溃疡/糜烂、肿块/息肉和出血的检出率分别为 69%、38%、46%和 17%。医生经验和解释时间并未显著影响检出率。
病理存在于少数几个帧中。评估的模式和速度有限。病变类型未通过手术或深部内镜方法证实。提供解释的读者相对较少。
总的来说，本研究中的检出率低于先前报道的水平，且不受经验增加的影响。与阅读 SingleView 25 相比，阅读 SingleView 15 和 QuadView 20 的检出率显著更高。将观察速度从 QuadView 20 增加到 QuadView 30 似乎对检出率没有显著影响。需要进一步研究质量控制措施来比较和提高病变检出率。
