Assessment of the diagnostic accuracy of T2*-weighted MR imaging for identifying hepatocellular carcinoma with liver explant correlation

Andrew D. Hardie, John Nance, Daniel J. Boulter, Michael K. Kizziah

Research output: Contribution to journalArticle

11 Scopus citations

Abstract

Background: T2*-weighted MRI may represent a novel method for identifying hepatocellular carcinoma (HCC). The goal of this study was to assess the diagnostic accuracy of T2*-weighted MRI for HCC with liver explant correlation. Materials and methods: A retrospective review identified 25 patients who had undergone liver transplantation with pre-operative T2*-weighted MRI. All patients had Child's-Pugh A (9), B (9), or C (7) liver disease with 13 transplanted for liver dysfunction and 12 for HCC. The T2*-weighted images were interpreted by 2 blinded, independent observers and the results compared with the explanted specimens. Sensitivity and specificity of T2*-weighted MRI for the identification of HCC was assessed. Results: By pathology, 16 HCC (mean largest diameter 2.1 cm; range 0.9-3.6 cm) were identified in 14 patients. Reader 1 had a sensitivity of 69% (95% confidence interval 41-88%) and a specificity of 100% (68-100%). Reader 2 had a sensitivity of 56% (31-79%) and a specificity of 100% (68-100%). There was a very good inter-observer agreement (kappa = 0.84). Conclusion: T2*-weighted MRI had a moderate sensitivity for identifying HCC but had an excellent specificity. A T2*-weighted MR sequence may be a useful component of a liver MRI protocol due to its high specificity for HCC, and may be particularly useful in patients unable to undergo gadolinium enhanced MRI.

Original languageEnglish (US)
JournalEuropean Journal of Radiology
Volume80
Issue number3
DOIs
StatePublished - Dec 1 2011

Keywords

  • Cirrhosis
  • Hepatocellular carcinoma
  • Iron
  • MRI

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging

Fingerprint Dive into the research topics of 'Assessment of the diagnostic accuracy of T2*-weighted MR imaging for identifying hepatocellular carcinoma with liver explant correlation'. Together they form a unique fingerprint.

Cite this