Multiparametric gadoxetic acid-enhanced MR versus dual-layer spectral detector CT for differentiating hepatocellular...

BackgroundIt can be difficult to differentiate hypervascular hepatocellular carcinoma (HCC) from hypervascular pseudolesion (HPL) such as arteriovenous shunts.PurposeTo determine retrospectively whether double-layer detector computed tomography (DLCT) can differentiate HCC from HPL compared to gadoxetate-enhanced magnetic resonance imaging (EOB-MRI).Material and MethodsWe retrospectively analyzed 46 patients who underwent EOB-MRI and DLCT for suspected HCCs. Arterial/portal phase and hepatobiliary phase (HBP) on EOB-MRI, T2-weighted (T2W) imaging, diffusion-weighted imaging (DWI), apparent diffusion coefficient (ADC), CT value, iodine-density (ID), atomic-number (Zeff), and electron-density (ED) of the lesion and liver were evaluated. The reduction rates of ID (R-ID) between each phase of the arterial/portal phase on EOB-MRI were calculated. ROC analysis was performed to determine the accuracy for differentiating HCC from HPL.ResultsThere were 55 HCCs and 14 HPLs. On DWI, 42, 11, and two HCCs showed high, slightly high, and iso intensity, respectively. However, all 14 HPLs showed iso intensity on DWI. Area under ROC curve (AUC) of DWI (0.982, 95% confidence interval [CI]=0.957-1) was significantly higher than that of HBP (AUC=0.714; 95% CI=0.580-0.849; P < 0.001), R-ID (AUC=0.742, 95% CI=0.580-0.903; P = 0.004), and ED of portal phase (AUC=0.786, 95% CI=0.640-0.891; P = 0.001) in differentiating HCC and HPL. ADC (<0.001), T2W imaging (<0.001), HBP (<0.001), ED-arterial-phase (<0.001), ED-portal-phase (=0.003), ED-equilibrium-phase (=0.001), R-ID-between-arterial/equilibrium-phase (=0.032), and R-ID-between-portal/equilibrium-phase (=0.042) showed significant differences between HPL and HCC.ConclusionDWI is most useful for differentiating HCC from HPL, although ADC, T2W, HBP, R-ID, and ED may also be relatively useful to differentiate between HPLs and HCCs.