Incorporating the hepatobiliary phase of gadobenate dimeglumine-enhanced MRI in the diagnosis of hepatocellular carcinoma: increasing the sensitivity without compromising specificity. (August 2016)

• Record Type: Journal Article
• Authors: Cortis, Kelvin; Liotta, Rosa; Miraglia, Roberto; Caruso, Settimo; Tuzzolino, Fabio; Luca, Angelo
• Is Part Of: Acta radiologica. Volume 57:Number 8(2016:Aug.)
• Page Start: 923
• Page End: 931
• Publication Date: 2016-08
• Subjects: Hepatocellular carcinoma (HCC) -- gadobenate dimeglumine -- magnetic resonance imaging (MRI) -- cirrhosis -- organ transplants; Radiology, Medical -- Periodicals; Radiography, Medical -- Periodicals; Radiotherapy -- Periodicals; 616.0757
• DOI: 10.1177/0284185115616291
• Languages: English
• ISSNs: 0284-1851

Abstract

Background: Diagnosis of hepatocellular carcinoma (HCC) is centered on wash-in of contrast during the arterial phase followed by washout during the portal or delayed venous phase. Nodules showing hypointensity on the hepatobiliary phase are also likely to represent HCC, however, the role of this phase is not yet established. Purpose: To investigate the role of the hepatobiliary phase on Gadobenate dimeglumine (Gd-BOPTA) magnetic resonance imaging (MRI) in characterizing HCCs lacking the typical arterial enhancement and venous washout. Material and Methods: Ninety-seven cirrhotic patients (78 men, 19 women; mean age, 58.5 years) who underwent liver transplantation (2004–2012) and Gd-BOPTA enhanced MRI within 3 months of surgery were retrospectively reviewed. A nodule-by-nodule analysis was performed, followed by liver explant correlation. Statistical analysis was then performed by a biostatistician using commercially available software. Results: A total of 193 HCCs were found in 97 liver explants, of which 24.9% (48/193) were not detectable on imaging. The 145 HCCs seen on imaging showed the typical wash-in/washout pattern (Pattern A) in 46.9% (68/145), arterial enhancement without washout (Pattern B) in 37.9% (55/145), and hypovascularity on arterial and venous sequences (Pattern C) in 15.2% (22/145). Pattern A was exclusive to HCC. Twenty-three of the 55 HCCs showing Pattern B were also hypointense on the hepatobiliary phase (Pattern B1). Combining Pattern B1 with Pattern A raises the sensitivity of HCC characterization from 46.9% to 62.8% ( P = 0.007), with no significant compromise on specificity. Conclusion: When coupled with Pattern A, Pattern B1 augments sensitivity of HCC characterization with no significant compromise on the specificity.