In a recent study, in which an unselected patient cohort and multiphase multi-detector row computed tomography (MDCT) were utilized, CEUS was shown to be significantly inferior to MDCT in the preoperative detection of hepatic metastases of colorectal cancer [26]. Generally, CT is a fast and relatively inexpensive imaging technique; its performance has been significantly improved with the development of helical CT and MDCT. This increases the speed, acquisition, and resolution, thus enabling imaging of the liver during various phases of contrast enhancement. During a single breath-hold, imaging of the entire breast and abdomen can take place, thereby eliminating respiratory motion artifacts [20]. On the other hand, MDCT is considered to be the imaging modality of choice in CLM. MDCT can generate slices of the liver with a thickness of less than 1 mm, subsequently enabling high-quality reformatted multi-planar (MPR) and volumetric three-dimensional (3D VR) reconstructions [21]. These reconstructions demonstrate the main features of the CLM such as size, margin, and relationship with the vascular and biliary structure, as well as the remaining liver volume.

Intravenous iodinated contrast media can be used to help characterize hepatic lesions based on their enhancement patterns during different phases of contrast circulation in the liver [3, 27]. Early in the arterial phase, the hepatic arterial anatomy can be visualized, which is useful for surgical planning, whereas hyper-vascular lesions tend to show more degrees of enhancement when compared to the surrounding hepatic parenchyma during the late arterial phase. Hypo-vascular lesions, such as CLM, become more pronounced during the portal-venous phase (Fig. 2.3). Through vascular reconstruction, the hepatic arterial and portal-venous anatomy can be visualized, thus eliminating the need for conventional angiography for the surgical planning [20].

In CT scanning with arterioportography (CTAP), the contrast agent is injected into either the superior mesenteric artery or the splenic artery via a percutaneously placed catheter. Only the liver parenchyma is contrast-enhanced (as in the portal phase), due to the fact that metastases are almost solely perfused by the hepatic artery. This technique is superior to other imaging techniques in the detection of lesions smaller than 2 cm, but the specificity is relatively low and false-positive findings should be identified. In practice, CTAP currently does not play a role in the diagnostic work-up of liver metastases.

The limitations of contrast-enhanced CT are the increased risk of radiation exposure (especially when several contrast phases are acquired), possible allergic reactions to the contrast media utilized, and low sensitivity for lesions smaller than 1 cm. Furthermore, the presence of fatty liver, which often occurs after chemotherapy, decreases the sensitivity and specificity of CT.

The FLR can be calculated by CT-imaging. With the help of software programs, the proposed line of resection can be drawn and the remaining liver volume calculated as a percentage of the whole liver volume. The regenerative capacity of the liver can be determined by the assessment of the hepatic response to portal vein embolization (PVE). The hepatic volume is measured via CT-imaging before and 3 weeks after PVE. The occurrence of a minimum of 5% hypertrophy of the liver within these 3 weeks is considered acceptable. In the case of reduced hypertrophy, the FLR should be larger in order to maintain adequate liver function after resection [4]. This could be the case in hepatic cirrhosis or chemotherapy-induced injury. In incidents of underlying hepatic disease, the hepatic function can be further determined via indocyanine green clearance or lidocaine conversion tests [28].

In many studies, MRI has been found to have the greatest sensitivity and specificity when compared to FDG-PET and CT [29]. MRI uses the vastly differing properties of water and fat to generate images based on the proportion of these in normal and pathological tissue. It can provide greater liver-to-lesion contrast without the need for radiation and detects smaller lesions due to the lower resolution of the images. With faster imaging sequences, like T1-weighted spoiled gradient echo and T2-weighted turbo spin echo, image acquisition of the entire liver can take place in a single breath-hold, thus limiting motion artifacts. In T1-weighted images, CLMs have low signal intensity, with a central area of even lower intensity (doughnut sign); these sequences are mainly useful for the assessment of parenchymal fatty infiltration (Fig. 2.4a). In T2-images, CLMs have intermediate to high signal intensity, making these sequences useful for the differentiation between solid and non-solid lesions (Fig. 2.4b).

The addition of diffusion-weighted imaging (DWI) to MRI improves the detection of CLM smaller than 1 cm, especially in unenhanced MRI examinations. DWI prevalently increases specificity/negative predictive value in gadoxetate disodium (Gd-EOB-DTPA)-enhanced examinations [30]. The technique uses the Brownian motion of water in tissues: the random movement of water which is modified by numerous interactions with intracellular organelles, cellular membranes, and macromolecules. Any architectural changes in tissues, such as the increased cellularity commonly seen in tumor tissue, can restrict the Brownian motion [31]. These changes in motion can be quantified by the calculation of the apparent diffusion coefficient (ADC), where the ADC is inversely correlated to the tumor cellularity [21] (Fig. 2.4d). Nowadays, these sequences are routinely used for MRI of the liver in many centers.

Contrast-enhanced MRI is equally as sensitive as CTAP, however, it has the advantage of the occurrence of fewer false positive findings [20]. Two types of contrast media are used in MRI: extra-cellular media and tissue-specific media. The most often used extra-cellular media are paramagnetic chelates of gadolinium. In the delayed arterial phase, CLMs have a cauliflower-like appearance with intense peripheral enhancement. Gadolinium-based contrast media are excreted by the kidneys, and may cause acute renal reactions such as contrast-induced nephropathy or (on rare occasions) nephrogenic systemic fibrosis, especially when macrocyclic gadolinium-chelates are used [32]. Therefore, the glomerular filtration rate should be taken into account when considering the administration of gadolinium contrast media [21]. Guidelines provided by the European Society of Urogenital Radiology are readily available [33].

The tissue-specific contrast media, super-paramagnetic iron oxide (SPIO), is selectively absorbed by the reticulo-endothelial system of the normal liver parenchyma, spleen, and lymph nodes. There is a shortening of the T2 relaxation time, thus decreasing signal intensity in the liver parenchyma. Since malignant lesions almost completely lack a reticulo-endothelial system, the CLMs appear as high signal lesions in T2-weighted images. This is a highly sensitive method for small focal liver lesions [27, 34, 35]. The infusion time of SPIO is relatively long (30 min), which in turn prolongs the study time and is associated with side-effects such as lower back pain and hypotension. Given this, the use of SPIO for liver imaging has decreased in popularity during the last years.

Godobenate dimeglumine (gd-BOPTA) is a contrast medium which combines the properties of an extra-cellular agent and a tissue-specific agent, therefore making both dynamic imaging and delayed phase imaging possible. Whereas dynamic phase imaging is important for lesion characterization, delayed phase imaging can increase the sensitivity. Even though sensitivity may not be as high as SPIO-MRI, the tumor characterization is more specific, which is especially important in cases with both benign and malignant lesions [20].

Another contrast agent is gadoxetate disodium (Gd-EOB-DTPA), which makes a comprehensive evaluation of the liver with the acquisition of both dynamic and hepatocyte phase images possible. This can potentially provide additional information that could be useful in the detection and characterization of small liver lesions [36]. A growing number of articles in the literature have demonstrated the usefulness of the hepatobiliary-specific MRI contrast agent, Gd-EOB-DTPA, in liver imaging. When using Gd-EOB-DTPA, there is no contrast-agent uptake in the liver-specific late phase of the metastases; instead, they appear distinctly hypo-intense and thus often enable reliable delineation from normal liver parenchyma [37]. In a recent study, significantly more patients with CLM in the Gd-EOB-DTPA-MRI group were considered to be eligible for surgery (39.3% vs 31.0%), and 26.7% for MRI with standard extra-cellular contrast-media and contrast-enhanced MDCT [38]. MRI has low sensitivity for extra-hepatic lesions, for example in the peritoneum and chest; it could therefore be used for the evaluation of small, unclear liver lesions detected on CT or in patients with allergies to iodinated contrast agents. MRI requires relatively more compliance from the patient, such as commands for inhalation and longer periods of lying still. Patients who are unable to hold their breath for more than 15 s are usually poor candidates for MRI (Fig. 2.5).

FDG-PET uses the increased glucose metabolism in tumor cells (the Warburg effect), where a radioactive tracer is accumulated in cells with increased hypermetabolism. Since metabolic abnormalities usually precede anatomical changes in malignant tumors, FDG-PET improves early detection of CLM. Furthermore, FDG-PET can be used to search the entire body for the presence of early signs of extra-hepatic disease such as peritoneal metastases and lymph node involvement before liver surgery [34]. In a meta-analysis, the use of FDG-PET was shown to affect clinical management through the detection of additional liver metastases and/or extra-hepatic disease, mostly resulting in the switch from the intended curative surgery to a palliative treatment course. Nonetheless, the survival rate tended not to differ between the patients selected for surgery with and without FDG-PET [39] (Fig. 2.6).