Imaging Techniques for the Diagnosis of Hepatocellular Carcinoma: A Systematic Review and Meta-analysis.

Abstract Several imaging modalities are available for diagnosis of hepatocellular carcinoma (HCC). To evaluate the test performance of imaging modalities for HCC. MEDLINE (1998 to December 2014), the Cochrane Library Database, Scopus, and reference lists. Studies on test performance of ultrasonography, computed tomography (CT), or magnetic resonance imaging (MRI). One investigator abstracted data, and a second investigator confirmed them; 2 investigators independently assessed study quality and strength of evidence. Few studies have evaluated imaging for HCC in surveillance settings. In nonsurveillance settings, sensitivity for detection of HCC lesions was lower for ultrasonography without contrast than for CT or MRI (pooled difference based on direct comparisons, 0.11 to 0.22), and MRI was associated with higher sensitivity than CT (pooled difference, 0.09 [95% CI, 0.07 to 12]). For evaluation of focal liver lesions, there were no clear differences in sensitivity among ultrasonography with contrast, CT, and MRI. Specificity was generally 0.85 or higher across imaging modalities, but this item was not reported in many studies. Factors associated with lower sensitivity included use of an explanted liver reference standard, and smaller or more well-differentiated HCC lesions. For MRI, sensitivity was slightly higher for hepatic-specific than nonspecific contrast agents. Only English-language articles were included, there was statistical heterogeneity in pooled analyses, and costs were not assessed. Most studies were conducted in Asia and had methodological limitations. CT and MRI are associated with higher sensitivity than ultrasonography without contrast for detection of HCC; sensitivity was higher for MRI than CT. For evaluation of focal liver lesions, the sensitivities of ultrasonography with contrast, CT, and MRI for HCC are similar. Agency for Healthcare Research and Quality. ( CRD42014007016).