Though capsule endoscopy has become the first-line noninvasive examination of the small bowel and seems to be an easily performed procedure, most overlook the importance and intensity of examining and interpreting the wirelessly obtained images. A standardized step-by-step process is crucial for an effective and accurate reading. First, an overview of the study is obtained including completeness, cleanliness, and search of areas with large amounts of fresh blood. After setting the anatomic landmarks, images are carefully viewed while adapting the reading speed. The reading process is time-consuming, can be tedious, and needs a lot of experience. The rate of images must be carefully adapted to the speed of the capsule unless lesions are going to be missed. Virtual chromoendoscopy may help in the differentiation of lesion types but not in their detection. All relevant lesions are recorded with a thumbnail. During the next step, thumbnailed images are classified according to standard terminology. This chapter provides hints for the interpretation of difficult images. An endoscopic diagnosis is made based on the type, number, localization, and distribution of lesions together with available clinical information. Scoring systems have been developed for further specification. Finally, a report is provided summarizing the information and giving a recommendation for further management of the patient. Artificial intelligence is emerging and increasingly incorporated into software. These tools can help shorten the reading time and potentially to reduce the miss rate of lesions. However, presently they cannot replace the human reader.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Evaluation of Small Bowel Capsule Endoscopic Images

  • Daniela Mueller-Gerbes,
  • Martin Keuchel

摘要

Though capsule endoscopy has become the first-line noninvasive examination of the small bowel and seems to be an easily performed procedure, most overlook the importance and intensity of examining and interpreting the wirelessly obtained images. A standardized step-by-step process is crucial for an effective and accurate reading. First, an overview of the study is obtained including completeness, cleanliness, and search of areas with large amounts of fresh blood. After setting the anatomic landmarks, images are carefully viewed while adapting the reading speed. The reading process is time-consuming, can be tedious, and needs a lot of experience. The rate of images must be carefully adapted to the speed of the capsule unless lesions are going to be missed. Virtual chromoendoscopy may help in the differentiation of lesion types but not in their detection. All relevant lesions are recorded with a thumbnail. During the next step, thumbnailed images are classified according to standard terminology. This chapter provides hints for the interpretation of difficult images. An endoscopic diagnosis is made based on the type, number, localization, and distribution of lesions together with available clinical information. Scoring systems have been developed for further specification. Finally, a report is provided summarizing the information and giving a recommendation for further management of the patient. Artificial intelligence is emerging and increasingly incorporated into software. These tools can help shorten the reading time and potentially to reduce the miss rate of lesions. However, presently they cannot replace the human reader.