Advanced Imaging Modalities for Hepatocellular Carcinoma: Is MRI with EOVIST Really Better?

Background: Hepatocellular cancer (HCC) is the third leading cause of cancer-related death worldwide and the third most common indication for liver transplantation in the United States. Efforts toward perfecting imaging-based diagnosis have increased to avoid the need for liver biopsy. Eovist (gadolinium- EOB-DTPA), compared to conventional gadolinium-enhanced MRI (MRI-Gd) or triple-phase contrast-enhanced computed tomography (CT), is considered a superior method for detection of hepatocellular cancer (HCC). In patients with cirrhosis, Eovist enhances lesion-to-liver contrast and differentiates vascular shunts and dysplastic nodules from HCC, an important distinction as outcomes of transplantation depend on the degree of cancer burden. We investigate whether MRI with Eovist (MRI-E) is more accurate for evaluation of HCC than MRI-Gd or CT. Methods: Retrospective analysis of all patients with HCC undergoing liver transplantation at Cedars-Sinai Medical Center from 2009-2014 was conducted. Multicentric tumors were included if they could be uniquely identified across modalities based on anatomic location. Number and size of lesions measured by MRI-E, MRI-Gd, or CT were compared to explant pathology using repeated measures ANOVA and linear regression analysis. Viability on imaging vs. pathology was compared using chi-squared tests. Results: Sixty-four patients with 137 HCC tumors were imaged with MRI-E (n=96), MRI-Gd (n=63), and/or CT (n=53); 33 tumors were measured with all 3 modalities. The number of lesions identified by MRI-E was highly concordant with pathology and higher than the number detected by MRI-Gd or CT (p<0.05). All three imaging modalities underestimated maximum tumor diameter relative to pathology (p=.0003). Maximum tumor diameter by MRI-Gd had stronger correlation with pathology than MRI-E or CT (p=0.008). MRI-E (χ2=3.52, p=0.061) and CT (χ2=3.57, p=0.059) were better at assessing viability than MRI-Gd (χ2=1.22, p=0.268). Conclusions: This is the first study to compare imaging of HCC using MRI-E, MRI-Gd, or CT to explant pathology. MRI with Eovist is a useful adjunct for liver transplant candidacy evaluation with superior assessment of the number of HCC lesions, but it may have limited precision when assessing lesion size.