Usefulness of three-dimensional, multidetector row CT (virtual gastroscopy and multiplanar reconstruction) in the evaluation of gastric cancer: a comparison with conventional endoscopy, EUS, and histopathology

Background. The aim of this study was to assess the role of three-dimensional multidetector row CT for detection, precise localization, and staging of gastric cancer by comparison with conventional endoscopy, EUS, and histopathology. Methods: Sixty-three patients with gastric cancer (31 early stage, 32 advanced) were evaluated by EGD with biopsies, EUS, and three-dimensional multidetector row CT between January 2003 to August 2003. Three-dimensional multidetector row CT findings were analyzed by a single radiologist blinded to the endoscopic findings. Among 63 patients, the findings were confirmed in 48 at surgery or by EMR. In the remaining cases, the findings were confirmed by EGD and biopsy specimen. The accuracy of three-dimensional multidetector row CT for detection, localization, and staging of gastric cancer was determined, compared with endoscopy, EUS, and histopathology. Results: Among the 63 patients, there were 67 gastric cancers. The overall accuracy of three-dimensional multidetector row CT for detection of gastric lesions was 94% (63/67), with accuracies of 96.7% (30/31) and 100% (32/32) for detection of, respectively, early and advanced stage gastric cancer. The overall accuracy, sensitivity, and specificity for EUS and three-dimensional multidetector row CT in the pre-operative determination of depth of invasion (T stage) were, respectively, 87.5%, 82.4%, and 96%; and 83.3%, 69.1%, and 94.4%. The accuracy, sensitivity, and specificity of EUS and three-dimensional multidetector row CT for lymph node staging was, respectively, 79.1%, 57%, and 89.5%; and 75%, 57.4%, and 89.3%. Conclusions: Three-dimensional multidetector row CT, along with virtual gastroscopy is a promising method for pre-operative evaluation of gastric cancer.