Evaluation of Gastric Lesion by Spiral CT: Comparison of Images by Different Water Ingestion Methods

Park, Seong Jin; Lee, Dong Ho; Ko, Young Tae

Abstract

PURPOSE
To determine the preferred method of drinking water before spiral CT scanning of a gastric lesion.
MATERIALS AND METHODS
Spiral CT scans of 92 patients with gastric lesion were obtained and evaluated. Patients drank tap water as oral contrast material and were scanned in the prone position; they were divided into two groups according to the method by which water was ingested. Group 1 patients drank 500ml 60 minutes before scanning; 500mL, 30 minutes before scanning; and 500mL just before scanning, while those in group 2 drank 800-1000mL just before scanning. In all cases, precontrast images were obtained, and an additional 300mL of water was given if the stomach was not adequately distended. Postcontrast images were obtained at 35 seconds, 80 seconds and 3 minutes after the start of infusion of contrast material. For 35- and 80-second scans, spiral CT was performed with 5-mm collimation, 7mm/sec table feed, and 5-mm reconstruction interval; for precontrast and 3-minute scans, 10-mm collimation, 10mm/sec table feed, and 10-mm reconstruction interval were used. There were 40 patients in group 1, and 52 in group 2. The images of the two groups were evaluated according to three criteria: gastric distension, mucosal enhancement, and tumor distinction. For each criterion, the images were evaluated by grading; an 'excellent' image scored 3, 'good' 2, 'ordinary' 1, and 'poor' 0. The differences in imaging quality between the two groups were statistically evaluated. Images obtained at 35 and 80 seconds, and at 3 minutes after the start of infusion of contrast material were compared, and the detectability of 18 confirmed cases of early gastric cancer was evaluated.
RESULTS
For gastric distension, the mean score was 1.65 in group 1, and 1.81 in group 2 (P=0.33); the corresponding figures for mucosal enhancement were 1.45 and 1.65 (P=0.11), and for tumor distinction, 1.30 and 1.52 (P=0.09). Between the two groups, there was therefore no statistical difference in image quality. With regard to postcontrast images, those obtained after a delay of 35 seconds were best; those obtained at 80 seconds were better than those obtained at 3 minutes. Fifteen of 18 case of early gastric cancer were detected on spiral CT, and the detection rate was 83.8%.
CONCLUSION
No statistically significant difference was seen between the two groups, and a single drink of water-just before scanning- is thus preferable to several drinks. Two scans, with images delayed for 35 and 80 seconds after contrast enhancement, are adequate. The use of these methods may improve the image quality of gastric lesions and the detection rate of early gastric cancer.