Diagnostic performance of computed tomography and magnetic resonance imaging for detecting peritoneal metastases: systematic review and meta-analysis

Primary end point was to assess diagnostic accuracy of computed tomography (CT) and magnetic resonance imaging (MRI) in detecting peritoneal metastases (PM). Secondary end points were determining the diagnostic sensitivity and specificity of CT in detecting PM according to the peritoneal cancer index (PCI), investigating correlations between radiological and surgical PCI, and comparing diagnostic yield of CT versus positron emission tomography (PET)/CT. We searched MEDLINE, Cochrane Library, Embase and Web of Science databases. Analytic methods were based on PRISMA. Pooled estimates for sensitivity, specificity, positive and negative likelihood ratios were calculated using fixed and random effect models. I (2) was used to evaluate heterogeneity. Of the 529 articles initially identified, 22 were selected for inclusion (934 patients). Cumulative data for per patient CT diagnostic accuracy were sensitivity 83 % (95 % CI 79-86 %), specificity 86 % (95 % CI 82-89 %), pooled positive LR 4.37 (2.58-7.41), and pooled negative LR 0.20 (0.11-0.35). On a per region basis CT performed best in epigastrium and pelvis. Correlation analysis showed a high correlation between CT-PCI and surgical-PCI scores, ranging from 0.49 to 0.96. MRI and PET/CT achieved similar per patient diagnostic accuracy. CT should be the preferred diagnostic imaging modality for detecting peritoneal metastases because of the robustness of the data. MRI and PET/CT should be considered second choices, until more consistent information on their diagnostic yield in detecting PM are obtained.