Chen Sheng1, Li Li, Wang Pei. 1. College of Information, Shanghai Normal University, Shanghai, People's Republic of China. chnshn@hotmail.com
Abstract
BACKGROUND: This paper presents a computer-aided method for automatic detection of the positioning of endotracheal, feeding and nasogastric tubes, and the identification of tube types in radiography for intensive care unit (ICU) patients. Application of this method may allow clinicians to detect the tube tips more easily and accurately, and thus improve the quality of patient care in the ICU. METHODS: One-hundred-and-seven portable X-ray images were collected from 20 patients, using a Kodak computed radiography system. It was determined whether each image did or did not have a tube and which kind of tube was present. In order to evaluate the performance of the proposed tube detection method, an experienced chest radiologist reviewed all images from the 20 patients and provided the true position of these tubes. The automatic detection results could then be compared with the actual results to determine the success rate. RESULTS: Preliminary results show that the computer-aided technique has a detection rate of 94% for endotracheal tubes, with 0.6 false positives per image, and 82% for both feeding and nasogastric tubes, with 0.4 and 0.5 false positives per image, respectively. CONCLUSION: The novel detection technique can accurately detect the tubes in ICU chest radiographs at a high sensitivity level with an acceptable false positive rate.
BACKGROUND: This paper presents a computer-aided method for automatic detection of the positioning of endotracheal, feeding and nasogastric tubes, and the identification of tube types in radiography for intensive care unit (ICU) patients. Application of this method may allow clinicians to detect the tube tips more easily and accurately, and thus improve the quality of patient care in the ICU. METHODS: One-hundred-and-seven portable X-ray images were collected from 20 patients, using a Kodak computed radiography system. It was determined whether each image did or did not have a tube and which kind of tube was present. In order to evaluate the performance of the proposed tube detection method, an experienced chest radiologist reviewed all images from the 20 patients and provided the true position of these tubes. The automatic detection results could then be compared with the actual results to determine the success rate. RESULTS: Preliminary results show that the computer-aided technique has a detection rate of 94% for endotracheal tubes, with 0.6 false positives per image, and 82% for both feeding and nasogastric tubes, with 0.4 and 0.5 false positives per image, respectively. CONCLUSION: The novel detection technique can accurately detect the tubes in ICU chest radiographs at a high sensitivity level with an acceptable false positive rate.