Background Surgery is the fundamental curative option for gastric cancer patients. Imaging scans are routinely prescribed in an attempt to stage the disease prior to surgery. Consequently, the correlation between radiology exams and pathology is crucial for appropriate treatment planning. Methods Systematic searches were conducted using Medline, Embase, and the Cochrane Central Register of Controlled Trials from January 1, 1998 to December 1, 2009. We calculated the accuracy, overstaging rate, understaging rate, Kappa statistic, sensitivity, and specificity for abdominal ultrasound (AUS), computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography (PET) with respect to the gold standard (pathology). We also compared the performance of CT by detector number and image type. A meta-analysis was performed. Results For pre-operative T staging MRI scans had better performance accuracy than CT and AUS; CT scanners using C4 detectors and multi-planar reformatted (MPR) images had higher staging performances than scanners with \4 detectors and axial images only. For pre-operative N staging PET had the lowest sensitivity, but the highest specificity among modalities; CT performance did not significantly differ by detector number or addition of MPR images. For pre-operative M staging performance did not significantly differ by modality, detector number, or MPR images. Conclusions The agreement between pre-operative TNM staging by imaging scans and post-operative staging by pathology is not perfect and may affect treatment decisions. Operator dependence and heterogeneity of data may Electronic supplementary material The online version of this article

Background Accurate preoperative staging is important in determining the appropriate treatment of gastric cancer. Recently, endoscopic ultrasound (EUS) has been introduced as a staging modality. However, reported test characteristics for EUS in gastric cancer vary. Our purpose in this study was to identify, synthesize, and evaluate findings from all articles on the performance of EUS in the preoperative staging of gastric cancer. Methods Electronic literature searches were conducted using Medline, Embase, and the Cochrane Central Register of Controlled Trials from 1 January 1998 to 1 December 2009. All search titles and abstracts were independently rated for relevance by a minimum of two reviewers. Metaanalysis for the performance of EUS was analyzed by calculating agreement (Kappa statistic), and pooled estimates of accuracy, sensitivity, and specificity for all EUS examinations, using histopathology as the reference standard. Subgroup analyses were also performed. Results Twenty-two articles met our inclusion criteria and were included in the review. EUS pooled accuracy for T staging was 75% with a moderate Kappa (0.52). EUS was most accurate for T3 disease, followed by T4, T1, and T2. EUS pooled accuracy for N staging was 64%, sensitivity was 74%, and specificity was 80%. There was significant heterogeneity between the included studies. Subgroup analyses found that annual EUS volume was not associated with EUS T and N staging accuracy (P = 0.836, 0.99, respectively). Conclusion EUS is a moderately accurate technique that seems to describe advanced T stage (T3 and T4) better than N or less advanced T stage. Stratifying by EUS annual volume did not affect EUS performance in staging gastric cancer.

Despite evolving preoperative imaging techniques, diagnostic laparoscopy continues to be of substantial value in staging patients with gastric cancer and in avoiding unnecessary laparotomy. The current data support DL for all patients with advanced gastric cancer.