Acoustic radiation force impulse imaging (ARFI) is a newly developed technique for the evaluation of tissue stiffness. It has been applied to evaluate liver fibrosis, but there is only limited data available for liver lesions characterization. In those preliminary data, ARFI measurements alone barely can differentiate benign and malignant liver nodules because of variations in stiffness of all types of masses, but cannot differentiate HCC from other malignancies [20, 21]. The substantial overlap between values of benign and malignant liver masses limits clinical usefulness of this technique.

On unenhanced CT scans, HCC often appears as a hypodense lesion. Typical HCC shows marked transient hyperenhancement compared to liver background during late arterial phase on dynamic studies. Typical HCC rapidly becomes iso- or hypoattenuating during portal or delayed venous phase [19, 22, 23]. Additional features on CT that may be associated to nodular HCC are the detection of a thin hyperenhancing tumoral capsule in portal or delayed venous phase [22, 24]; the presence of mosaic pattern, especially in large tumors, with different internal areas of differentiation[22, 24]; and the expansive vein thrombosis.

HCC has variable signal intensity on MRI, but predominantly it shows low signal intensity (SI) on T1-weighted imaging and high SI on T2-weighted imaging. Small HCCs may present fatty metamorphosis, causing hyperintense SI on T1-weighted images. Fat content can be confirmed when the SI of the lesion drops on opposed-phase imaging relative to in-phase imaging [25]. HCC can be difficult to detect on T2-weighted images because of heterogeneity of the cirrhotic liver, which obscures mildly hyperintense and isointense tumors.

On dynamic phases of gadolinium-based MRI, typical HCCs use to show hyperenhancement in the hepatic arterial phase compared to the surrounding liver parenchyma followed by hypointensity (washout) in the venous and/or delayed phases (Fig. 40.2). The fast decrease of SI on the venous and/or on delayed phases has been attributed to the loss of portal-venous supply at the latest stages of hepatocarcinogenesis [26, 27]. In a low proportion HCCs may be hypoenhancing and nearly isointense in arterial phase and persists with low SI during venous phases.

Rarely, HCCs may remain hyperintense relative to liver parenchyma on the delayed phases. This atypical vascular pattern had been described in scirrhous HCCs and use to be secondary to the presence of prominent central fibrous stroma [28].

MRI has been reported to be more sensitive than CT to depict the peritumoral fibrous capsule or pseudocapsule, which is a frequent finding on HCC. Both capsule and pseudocapsule appear on MRI as thin rim that surrounds the HCC, presenting hypersignal on T2 and hyperenhancement on delayed venous phase [29]. This fact should represent a helpful diagnostic sign in cases of atypical vascular pattern. However, the presence of pseudocapsule can be identified in a small proportion of atypical nodules that finally resulted HCCs [30].

Recent studies have shown that the detection of all types of focal liver lesions, including small HCCs, is improved with diffusion-weighted imaging (DWI) compared with conventional T2-weighted imaging [31, 32]. In highly cellular tissues such as HCC or other malignant lesions, diffusion of water protons is impeded within the lesion, resulting in high SI of the lesion versus background liver [33]. DWI is now a routine part of all liver MRI protocols and may improve the detection of HCC or reader’s confidence in the diagnosis of HCC. Similar to T2-weighted, hyperintense SI on DWI is not specific for HCC, even in the cirrhotic liver, and other focal liver lesions may show this appearance, such as intrahepatic cholangiocarcinoma (ICC) [34].

Recently, An et al. [35] reported that DWI together with dynamic MRI phases might be useful to predict the histological grade of HCC. The authors found a tendency towards the higher grades of HCC showing the more restriction on DWI. Also, the results from this study showed that characteristics from DWI and subtraction of unenhanced images from arterial phase images can be used to predict the HCC histological grade, especially when SI characteristics on both MRI sequences are considered together: poorly differentiated HCC showed higher SI on DWI and/or arterial enhancement on subtraction imaging [35] (Fig. 40.3). In spite of these are novel results, the impact of histological grade of HCC over the decision-making process is still under discussion.

An important technical aspect to be considered for interpreting the findings on DWI is that the sensitivity for HCC detection on DWI decreases with increasing severity of cirrhosis since the SI of the liver background increases and the liver to HCC contrast decreases [36]. As a general rule, the apparent diffusion coefficient (ADC) quantification of malignant liver lesions, including HCC, is lower than benign lesions [37], but the overlap on ADC values found among different liver malignancies and premalignant conditions, such as dysplastic nodules, and the heterogeneity among different studies regarding the technical characteristics on DWI sequences represent a major limitation at the time to establish a threshold to define the ADC value for HCC characterization [33, 37].

One of the major changes in liver imaging in the past years is the introduction of gadolinium-based contrast agents with liver-specific properties including gadobenate dimeglumine and gadoxetic acid. Due to the higher biliary excretion of gadoxetic acid (50 %) compared to gadobenate dimeglumine (3–5 %), gadoxetic acid is more often reported in the literature. Both contrast media provide information regarding tumor vascularity during the arterial and portal-venous phases of the study and functional information about the presence or absence of functional hepatocytes during the hepatobiliary (HB) phases generally obtained 20 min after gadoxetic acid injection.

The ultimate factor that conditions the gadoxetic acid uptake during the HB phase is the expression of two different transport systems located at the sinusoidal and canalicular membranes of the cell, including the OATP1B1 and OATP1B3, a solute carrier transporter, and the multidrug resistance protein 2 (MRP2), a canalicular transporter [38]. Therefore, non-hepatocyte focal lesions or those with nonfunctional hepatocytes, such as most HCCs, will not accumulate contrast, resulting in low-signal-intensity foci relative to the high SI enhancing liver parenchyma in the HB phase (Fig. 40.4). The percentage of the gadoxetic acid that is not cleared by the hepatobiliary system is excreted by glomerular filtration in the kidneys. Therefore, the liver uptake on HB is surrogated by the renal and liver function [39]. Even though several authors have attempted to find biological markers to predict a qualitative enhancement on the HB phase, including indocyanine green, prothrombin, albumin, liver enzymes, and Child-Pugh score, there is no clear biological marker that is consistently associated with reduced uptake of liver parenchyma on the HB phase. Nevertheless, the general trend shows reduced liver enhancement on the HB phase with increasing severity of cirrhosis, but not all patients with severe cirrhosis (Child B and C) had reduced uptake of contrast material [39–42].

The clinical recommended dose of gadoxetic acid is 0.025 mmol/kg instead of the 0.1 mmol/kg that is the recommended dose for the other nonspecific extracellular gadolinium-based contrast agents. This low clinical dose of gadoxetic acid may lead to lower enhancement of focal liver hypervascular lesions and portal vein during the arterial and venous phases, respectively, as compared to nonspecific gadolinium, hampering the detection and characterization of HCC and the assessment of arterial anatomy and portal vein patency [43]. Moreover, the injection volume of gadoxetic acid is also smaller, resulting in a possible acquisition timing error and truncation artifacts in the arterial phase. Doubling the dose of gadoxetic acid has also been suggested, as the degree of focal hepatic lesion enhancement in the arterial and parenchymal liver uptake in HB phases is dose dependent. In a study of patients with cirrhosis and HCC, it has been shown that doubling the dose improves the tumor-to-liver contrast during the arterial phase in all patients and also during the HB phase [44]. This may be especially relevant in the subgroup of patients in whom the liver function impairment causes a lower liver uptake of liver-specific contrast agent on HB phase, such as in Child B or C cirrhotic patients. Nevertheless, the weak uptake in the HB phase may mask the lesions as they appear similar to background liver parenchyma, canceling the potential contribution of the HB phase over the lesion characterization [44, 45]. The clinical impact of this policy in terms of lesion detection and the cost-effective analysis had not been evaluated.

There is more limited experience in clinical practice using gadoxetic acid in comparison with extracellular nonspecific gadolinium. The former is well tolerated and has a rate of adverse events similar than those reported for nonspecific gadolinium chelates. Nevertheless, the bolus injection of gadoxetic acid had been associated to an acute self-limiting dyspnea occurring significantly more often than when intravenous unspecific extracellular gadolinium is used. The higher rate of this side effect results in a more frequent degradation of the quality of arterial phase, limiting the detection of arterial-enhancing nodules [46]. In order to avoid a higher rate of these artifacts, it is advisable to inject gadoxetic acid bolus at a rate not higher than 1 mL/s [47].

There is very limited data on MR elastography as a tool to characterize focal liver lesions. MR elastography is an imaging modality allowing direct visualization and quantitative measurement of propagating mechanical shear waves in biological tissue. The technique is used to obtain spatial maps (elastograms) and measurements of shear wave displacement. Preliminary data suggests that the technique can differentiate benign solid hepatic focal lesions (hemangiomas, focal nodular hyperplasia, adenoma) from malignant focal lesions (metastasis, cholangiocarcinoma, HCC), but the important overlap of stiffness values observed between different malignant lesions does not allow their proper characterization [48].