Diagnostic Accuracy of Imaging Modalities in the Evaluation of Vascular Invasion in Pancreatic Adenocarcinoma: A Meta-Analysis

BACKGROUND: The extent of vascular invasion is a key factor determining the resectability of non-metastatic pancreatic adenocarcinoma. The purpose of this study is to determine the diagnostic accuracy of computed tomography (CT), endoscopic ultrasound (EUS), and magnetic resonance imaging (MRI) in the pre-operative evaluation of vascular invasion in pancreatic adenocarcinoma, with surgery as the reference standard. METHODS: A search of the MEDLINE database for relevant articles in the English language published between January 2000 and February 2009 was performed. From each study, 2 × 2 tables were obtained, and pooled sensitivity, specificity, positive likelihood ratios, negative likelihood ratios and diagnostic odds ratios were calculated for each modality, along with a summary receiver operating characteristics (SROC) curve. RESULTS: 16 studies with a total of 797 patients who had surgical assessment of vascular invasion were included in the analysis. Several studies evaluated more than one imaging modality, allowing 24 datasets to be obtained in total. Sensitivity was highest for CT (0.73, 95% CI 0.67 - 0.79), followed by EUS (0.66, 95% CI 0.56 - 0.75) and MRI (0.63, 95% CI 0.48 - 0.77). The specificity for all three imaging modalities was comparable. The diagnostic odds ratios for CT, EUS and MRI were 45.9 (95% CI 18.0 - 117.4), 23.0 (95%CI 9.4 - 56.6), 23.9 (95% CI 5.4 - 105.1) respectively. CONCLUSION: CT was more accurate than EUS and MRI in the evaluation of vascular invasion in pancreatic adenocarcinoma and should be the first line investigation in pre-operative staging.