Computed tomography for the diagnosis of varices in liver cirrhosis: a systematic review and meta-analysis of observational studies

Objectives: This systematic review and meta-analysis aimed to evaluate the diagnostic accuracy of contrast-enhanced computed tomography (CT) for varices in liver cirrhosis.

Methods: PubMed and EMBASE databases were searched for the literature identification. The area under the summary receiver operating characteristic curve (AUSROC), sensitivity, specificity, positive and negative likelihood ratio (PLR and NLR), and diagnostic odds ratio (DOR) were calculated. We performed the subgroup analyses according to the location of varices, CT technique, and study design. The study quality was assessed according to the QUADAS-2 tool.

Results: Seventeen papers were eligible. The study quality was modest. The AUSROC was 0.8975 and 0.9494 for predicting any size and high-risk varices, respectively. Summary sensitivity, specificity, PLR, NLR, and DOR of CT for predicting any size and high-risk varices were 0.87/0.80/3.67/0.18/22.70 and 0.87/0.88/7.52/0.12/65.55, respectively. According to the location of varices, the AUSROC was 0.9127 for predicting any size gastric varices alone; and the AUSROC was 0.8958 and 0.9461 for predicting any size and high-risk esophageal varices alone, respectively. According to the CT technique, the AUSROC of multi-detector CT (MDCT) was 0.9047 and 0.9490 for predicting any size and high-risk varices, respectively; and the AUSROC of MDCT esophagograms for predicting any size and high-risk varices was 0.8735 and 0.9664, respectively. In the subgroup analysis of prospective studies, the AUSROC was 0.9122 and 0.9507 for predicting any size and high-risk varices, respectively.

Conclusion: CT had a high accuracy for the diagnosis of varices in liver cirrhosis.