Diagnostic and Interventional Imaging (2017) 98, 379—391

REVIEW /Abdominal imaging

Diagnosis of hepatocellular carcinoma: An

update on international guidelines
C. Cassinotto a,∗, C. Aubé b, A. Dohan c

a
Department of diagnostic and interventional imaging, Hôpital Haut-Lévêque, university
hospital of Bordeaux, CHU de Bordeaux, 1, avenue de Magellan, 33604 Pessac cedex, France
b
Department of diagnostic and interventional imaging, university hospital of Angers, 49933
Angers, France
c
McGill university health center, department of radiology, McGill university health center,
Montreal, QC, Canada

KEYWORDS Abstract Imaging is essential for the successful management of patients with or at risk of
Hepatocellular developing hepatocellular carcinoma (HCC). If ultrasound remains the key screening modality,
carcinoma; computed tomography and magnetic resonance imaging (MRI) can play a major role in the
Cirrhosis; characterization and noninvasive diagnosis of nodules in patients at risk of developing HCC.
Noninvasive Each technique has succeeded in adapting to the wide histological spectrum of focal liver
diagnosis; lesions. In this review, we discuss recent advancements in imaging techniques and evaluation
Magnetic resonance — notably diffusion-weighted imaging, contrast-enhanced ultrasound, and liver-specific MRI
imaging; contrast agents — as well as their addition to international guidelines and reporting systems
Contrast-enhanced such as the Liver imaging reporting and data system (LI-RADS).
ultrasound © 2017 Published by Elsevier Masson SAS on behalf of Editions françaises de radiologie.

In France, as in many countries worldwide, the incidence of primary liver cancer has
significantly increased since the 1980s [1]. In the United States, a study that estimated
the incidence of cancer and associated mortality rates in upcoming years reported that
primary liver cancer is expected to become the third largest cause of death by cancer by
2030, behind lung and pancreatic cancer but in front of bowel cancer [2]. On one hand, this
increase in the incidence of liver cancer may be attributed to an increase in the incidence
of chronic liver disease particularly alcohol-related liver disease and liver steatosis in
Western countries but, on the other hand, this increase may be due to improvements in

∗ Corresponding author.
E-mail address: [email protected] (C. Cassinotto).

http://dx.doi.org/10.1016/j.diii.2017.01.014
2211-5684/© 2017 Published by Elsevier Masson SAS on behalf of Editions françaises de radiologie.
380 C. Cassinotto et al.

cirrhosis patient care, which gives hepatocellular carcinoma developing into malignant HCC — or occur as de novo HCC.
(HCC) the extra time it needs to appear and develop [3,4]. Following the notable initiative of the Japanese pathologist
Particularly in patients with liver steatosis, HCC does not M. Kojiro, the nomenclature related to hepatocarcinogen-
develop after several years of progressive cirrhosis, but esis was updated and improved in 2009 [18]. This resulted
rather tends to appear during the preclinical stage, which in harmonized and more specific pathological criteria for
raises many questions about our current screening strategies defining the different types of precancerous hepatocellu-
[5,6]. lar nodules (also known as precursor lesions) and malignant
Over the last 15 years, imaging has played an ever- lesions (Fig. 1) [19]:
growing role in the diagnosis and staging of HCC. Radiologists • cirrhotic nodules (formerly referred to as regenerative
are involved in many of the key steps of HCC patient man- nodules): these nodules typically arise in the setting
agement, including assessment of the severity of chronic of cirrhosis — usually by the thousands — and appear
liver disease using noninvasive techniques [7—9], screening small, well-defined, and surrounded by an even layer of
of patients at risk of developing HCC, noninvasive diag- peripheral fibrous scar tissue. Microscopic dysplastic foci
nosis of HCC and other nodules that can develop in sometimes arise within these cirrhotic nodules and are
cirrhotic livers [10], and patient surveillance. Interventional thought to progressively develop into dysplastic nodules;
radiology is another important aspect of patient manage- • dysplastic nodules: generally with a diameter of
ment: radiologists obtain samples for pathological analysis 1—1.5 cm, these nodules are classified as low-grade dys-
via image-guided liver biopsy and perform both palliative plastic nodules (L-DN) or high-grade dysplastic nodules
and curative percutaneous and endovascular procedures (H-DN) based on the degree of cytologic atypia and archi-
[11—13]. tectural changes. Histologically, L-DNs tend to appear
Knowledge in the field of HCC imaging is growing rapidly different from regenerative nodules, and are therefore
and has prompted international medical societies to regu- considered precancerous lesions with a slightly higher risk
larly update their guidelines. In this review, we discuss the of becoming malignant. On the other hand, the histolog-
main achievements in the field of HCC imaging over recent ical features of H — DNs are similar to those of well —
years, as well as the consequences of imaging advances on differentiated HCCs, though they are differentiated based
different international guidelines and reporting systems. on a lack of stromal invasion. H-DNs and should be con-
sidered as advanced HCC precursors with a high risk of
malignant transformation;
Hepatocarcinogenesis • early HCC: early HCC tumors are vaguely nodular, gener-
HCC is a complex multistage process in which liver cells ally measure < 2 cm, and are typically differentiated from
repeatedly accumulate epigenetic and genetic damage that H-DNs by stromal invasion, which is defined as tumor cell
progressively gives rise to populations of precancerous cells invasion into the portal tracts or fibrous septa. The inva-
that then undergo malignant transformation [14—20]. This sive potential of early HCC lesions is low, and they do not
process can either occur gradually over time — starting with display the typical histological features of tumor aggres-
the appearance of hyperplasic foci that progress into dys- siveness such as the presence of a tumor capsule, satellite
plastic nodules and then early-stage HCC before eventually nodules, or microvascular/macrovascular invasion;

Figure 1. Schema showing the progression of precursor nodules towards malignancy during hepatocarcinogenesis. DN: dysplastic nodule;
HCC: hepatocellular carcinoma; LCC: large liver cell change; SCC: small liver cell change.
Reprinted from Park [19] with permission from Archives of Pathology & Laboratory Medicine. Copyright 2011 College of American Pathologists.
Diagnosis of hepatocellular carcinoma: An update on international guidelines 381

• progressed HCC: these clearly malignant lesions show vas- with alcoholic cirrhosis was not reported and has still not
cular invasion and can metastasize. If early HCC is a been provided in the latest EASL and AASLD guidelines [6].
potential precursor of progressed HCC, the latter may also Nevertheless, alcohol consumption remains the main cause
occur de novo or develop from cancerous nodules that of cirrhosis in France. A recent report estimated the inci-
arise within H-DN. Progressed HCC can appear as either dence of HCC in patients with alcoholic cirrhosis to be 2.6%
small (< 2 cm) or large HCC (≥ 2 cm). Typically, progressed per year [22]. This value is identical to the incidence of HCC
HCC nodules are clearly delimited and encapsulated. The in patients with cirrhosis due to nonalcoholic steatohepatitis
risk of higher histological grading, vascular invasion, and [23]. In 2005, AASLD recommended that patient subgroups
metastasis increases with the size of the tumor; with chronic HBV infection (even without cirrhosis) and an
• multifocal HCC: the simultaneous occurrence of multiple additional risk factor (such as Asian or African ancestry or a
HCCs may reflect either the synchronous development of family history of HCC) be included in surveillance programs
independent HCCs (multicentric hepatocarcinogenesis) or [24]. The updated 2011 AASLD guidelines and joint 2012
intrahepatic metastasis by dissemination from a primary EASL/European organisation for research and treatment of
tumor. cancer (EORTC) guidelines are summarized in Fig. 2 [25,26].
Both series of guidelines place special emphasis on the sub-
group of patients with chronic HCV infection and severe
Screening for HCC
fibrosis (stage F3 of the METAVIR scoring system) [27]. Such
The guidelines for HCC-screening programs are regularly patients are already included in EASL-based surveillance
updated by two main international societies — the European programs and should also be included in the next edition
association for the study of the liver (EASL) and the American of the AASLD guidelines. In the same way, patients with
association for the study of liver diseases (AASLD) — based nonalcoholic steatohepatitis without cirrhosis were consid-
on HCC epidemiological data reported in the literature for ered for inclusion as surveillance targets due to the risk
various population subgroups. of HCC before the development of cirrhosis; however, the
benefits of surveillance have not been demonstrated in this
population [6,26]. It should also be noted that despite the
Target populations reduced risk of developing HCC, patients who have been
The first edition of the EASL guidelines, published in 2001, cured of their chronic viral infection should continue to be
stated that all patients with cirrhosis are at risk of devel- monitored.
oping HCC, and therefore recommended that all cirrhotic
patients be included in surveillance programs [21]. At that
time, the incidence of HCC was estimated to be 2.5% for Screening tools
patients with hepatitis B virus (HBV) cirrhosis, 3—8% for Liver ultrasound is the key modality for HCC screening, all
patients with hepatitis C virus (HCV) cirrhosis, approx- the more so now that monitoring serum alpha-fetoprotein
imately 5% for patients with hemochromatosis-induced levels is no longer recommended [25,26]. Ultrasound is
cirrhosis, and still unknown but thought to be of the same a screening modality that demonstrated adequate sen-
order of magnitude for other forms of cirrhosis such as sitivity (60—90%) and excellent specificity (90%) [28,29].
primary biliary cirrhosis, cirrhosis due to autoimmune hep- Ultrasound has the advantages of being noninvasive, well-
atitis, and Wilson’s disease. The incidence of HCC in patients accepted by the patient population, and relatively low-cost.

Figure 2. Description of the target population for HCC screening based on the European (EASL/EORTC) and American (AASLD) guidelines
[25,26].
382 C. Cassinotto et al.

However, ultrasound has several limitations, notably that Such noninvasive diagnostic features have the advantage
it is highly operator-dependent, its performance is lower of excellent specificity, even though sensitivity decreases
in obese patients due to the impaired acoustic window, with nodule size. Indeed, sensitivity as low as 33% has been
and its sensitivity for small nodules is limited (63% for described for cirrhotic liver nodules that measure < 2 cm
HCCs measuring < 2 cm) [30]. Ultrasound sensitivity has been [33]. Malignancy cannot be excluded in the absence of such
shown to greatly improve when performed by operators who features, and alternative modalities or a combination of
have been trained in or are specialists in imaging the liver other features should be used to characterize small nodules.
parenchyma [31]. With the rapid increase in the overall pub-
lic health impact of HCC, training ultrasound operators to
adequately screen for HCC will probably be a key challenge Contrast-enhanced ultrasound
in the coming years. In obese patients, using a slice-imaging
modality (e.g., computed tomography [CT] or magnetic res- Sidelined for a time
onance imaging [MRI]) because of the impaired acoustic
window in ultrasound is not recommended by U.S. guide- When the AASLD and EASL/EORTC guidelines were updated
lines and is considered questionable by European guidelines, in 2011 and 2012, respectively, the use of contrast-enhanced
mainly because of exposure to ionizing radiation when using ultrasound (CEUS) was not included in the decision algo-
CT and the significant increase in the cost of using MRI rithms. However, this technique has several advantages such
screening [25,26]. Finally, the recommended length of time as its sensitivity in detecting arterial-phase hypervascular-
between follow-up visits, which needs to take into account ity due to the real-time visualization of the contrast agent
the tumor growth and the incidence in the target population, in the arterial blood supply (as long as the nodule is visible
has been set at 6 months in both sets of guidelines [25,26]. on standard ultrasound). However, its main disadvantage for
screening is its lack of specificity. Indeed, due to its excel-
Basis for the radiological diagnosis of HCC lent sensitivity in detecting hypervascularity, CEUS revealed
that most secondary liver lesions, which are typically found
Hepatocarcinogenesis is accompanied by major vascu- to be hypovascular during the arterial phase on CT or
lar changes, including tumor neoangiogenesis and venous MRI, actually showed transient arterial hypervasculariza-
drainage that switches from hepatic veins to hepatic sinu- tion. In a study that used CEUS to assess 50 liver metastases
soids and then to portal veins. The blood flow to the (including 28 colorectal, 6 neuroendocrine, 6 pancreatic,
healthy liver and most regenerative nodules is mainly sup- and 6 melanoma metastases), 88% showed arterial-phase
plied by the portal vein (approximately 80%) and arterial hypervascularity [34]. In addition, it has been shown that
blood from the hepatic artery (approximately 20%). As the specificity of portal- or delayed-phase washout was only
dysplastic nodules gradually develop and progress to HCC slightly more specific because this feature is observed in
nodules, this ‘‘normal’’ combined portal/arterial blood sup- nearly all malignant lesions [35], as well as in a number
ply decreases and is replaced by an ‘‘abnormal’’ supply of benign lesions. In a study on 74 benign lesions (includ-
from arterial tumor vessels. This tumor arterial angiogen- ing 29 angiomas, 31 cases of focal nodular hyperplasia, and
esis results in the hypervascular features of HCC that are 7 adenomas), washout was observed in 36% of cases [36].
seen during the arterial phase. In the first edition of the Nevertheless, in daily practice, the observed hypervas-
EASL guidelines published in 2001, hypervascularity dur- cularity of liver metastases from primary colorectal cancer
ing the arterial phase following the injection of a contrast (the most frequent cause of liver metastasis) is weaker and
agent (in combination with lesion size and alpha-fetoprotein more transient than that of HCC. The main issue that led to
levels) was the primary radiological feature used to diag- the exclusion of CEUS from the guidelines was the possible
nose HCC [21]. However, the sensitivity of these features similarities between the radiological features of cholan-
was poor because they do not apply to lesions measur- giocarcinoma and HCC in patients with cirrhosis [37]. The
ing < 2 cm, and their specificity was also low. Indeed, in differences in the patterns seen between CEUS and slice-
a study that compared radiological diagnosis with patho- imaging modalities such as CT and MRI are due to the nature
logical data from liver explants, > 90% of hypervascular of the contrast agent used. The contrast agents routinely
lesions measuring < 2 cm were found to be not HCC but used in CT and MRI are called ‘‘extracellular’’ because once
instead a great variety of benign lesions or pseudolesions the contrast bolus has passed through the arteries, it diffuses
such as cirrhotic nodules, dysplastic nodules, arteriove- into the interstitial space, without cellular uptake, before
nous shunts, or perfusion issues; the development of these being excreted by the urinary system. However, the main
benign lesions was promoted by parenchymal rearrange- agents used in CEUS are strictly intravascular and consist
ments related to cirrhosis [32]. Furthermore, the decrease of microbubbles that can be detected with great sensitivity
in and then loss of the portal blood supply in combination using ultrasound; therefore, these agents remain within the
with the increase in flow within the new tumor vessels also blood vessels, arteries, capillaries, and veins and are not
results in the hypoenhancement of HCC nodules in compar- taken up by the cells and do not diffuse into the interstitial
ison with the liver parenchyma during the portal or delayed space.
phase. This apparent hypovascularity observed during the
portal and/or delayed phase is also called ‘‘washout’’. Since . . .towards gradual reinstatement
2005, updates to the guidelines have included these new,
more specific features for diagnosing HCC: hypervascular- Nevertheless, CEUS continues to play a role in charac-
ity during the arterial phase and washout during the portal terizing nodules in patients at risk of developing HCC
and/or delayed phase. and was recently included in the German, British, and
Diagnosis of hepatocellular carcinoma: An update on international guidelines 383

Italian guidelines (www.drg.de, www.nice.org.uk, and cholangiocarcinomas, which tend to show irregular, mildly
www.webaisf.org, respectively). Its sensitivity for detecting enhanced peripheral enhancement. These results revived
hypervascularity during the arterial phase is higher than that the use of CEUS in 2016 with the first edition of a specific
of CT or MRI. Still, the sensitivity of CEUS for diagnosing HCC Liver imaging reporting and data system (LI-RADS) — type
according to the EASL and AASLD criteria remains low, even algorithm for CEUS that was issued by the American college
if its specificity is high. Therefore, there are two different of radiology and called CEUS LI-RADS [42]. In this freely
opinions regarding the use of CEUS around the world: available online algorithm [42], the diagnosis of HCC (i.e.,
• in the decision algorithm published in 2014 by the Japan LR-5; see Section 6.2.1) can be confirmed in at-risk patients
society of hepatology, CEUS is used as a second intention with a nodule > 10 mm that shows typical enhancement
modality when a nodule does not demonstrate hypervas- kinetics, e.g., nonperipheral, nonglobular hypervascular-
cularity but does show other suspicious features (such as ization of the whole or part of the lesion associated with
a hypointense MRI signal during the hepatobiliary phase) moderate delayed-phase washout (> 60 seconds) (Fig. 3).
[38];
• some authors, such as Jang et al., promote using CEUS as a
first-line screening modality due to its very high specificity Hepatospecific MRI contrast agents
[39].
Gadoxetic acid
Several further radiological features that characterize
HCC using CEUS are worth noting. Several studies have Traditionally, the contrast agents used in CT or MRI are
shown that the washout of HCC nodules is relatively mild extracellular contrast agents. After distribution via the cir-
and occurs at a fairly late stage — e.g., > 60 seconds after culation, these agents diffuse into the extracellular space
the injection of the contrast material — which is not before being eliminated almost entirely through the kidneys.
seen in other malignant lesions such as metastases and Gadoxetic acid (or gadolinium ethoxybenzyl diethylenetri-
cholangiocarcinomas that demonstrate an earlier and more amine pentaacetic acid [Gd-EOB-DTPA] and also known as
pronounced washout [40,41]. Moreover, the arterial-phase primovist in Europe or eovist in the USA; Bayer Health-
enhancement of HCC seen on CEUS can actually be specific, Care, Berlin, Germany) — a hepatospecific contrast agent —
appearing either as the massive enhancement of the whole was approved for use in Europe in 2004 (except in France
lesion or the part of the lesion that is not predominantly where the product has not been marketed) and in the
located at the periphery and shows no discontinuities. These USA in 2008. This contrast agent is taken up by normal
features allow discrimination between HCC, the main benign functional hepatocytes before being excreted in approxi-
lesions (e.g., angioma or focal nodular hyperplasia), and mately equal proportions by the hepatobiliary system and

Figure 3. Summary of the CEUS LI-RADS classification [42]. In this decision algorithm, the diagnosis of HCC (LR-5) can be confirmed in at-risk
patients with a nodule measuring > 10 mm and showing typical enhancement kinetics, i.e., nonperipheral, nonglobular hypervascularization
of the whole or part of the lesion associated with a moderate delayed washout (> 60 seconds after administration).
384 C. Cassinotto et al.

the kidneys. During the arterial and portal phases, gadoxetic appears hypointense during the hepatobiliary phase is very
acid provides enhancement patterns that are very simi- suggestive of HCC. On the contrary, high signal intensity
lar to those obtained using extracellular contrast agents. during the hepatobiliary phase is strongly indicative of a
Gadoxetic acid’s advantage lies in the analysis of its uptake benign lesion [57]. Nevertheless, it is advisable to proceed
by hepatocytes during the delayed phase, which is gener- with caution: while imaging during the hepatobiliary phase
ally assessed 20 minutes after administration and called the considerably improves the sensitivity of detecting HCC, its
hepatobiliary phase. During the hepatobiliary phase, func- specificity remains limited. A hypointense signal observed
tional hepatocytes that have taken up gadoxetic acid are during the hepatobiliary phase does not have the same level
enhanced, whereas non-hepatocytes and tumor cells are not of specificity as washout during the portal phase and/or
enhanced and therefore appear hypointense. Within just a after 3—5 minutes, and therefore does not have the same
few years, several studies reported the added value of using diagnostic value. Many cirrhotic and L-DNs that are not
this hepatospecific contrast agent. By improving the detec- observed on other sequences will appear with hypointense
tion of metastases during preoperative work-ups [43—45], signals during the hepatobiliary phase. This is notably the
gadoxetic acid has also been shown to be a promising option case for nodules measuring > 2 cm, for which the specificity
for assessing liver function [46,47] and now plays in impor- of hepatobiliary-phase images for diagnosing HCC has been
tant role in characterizing benign and malignant liver tumors reported to reach only 33% [60]. Cholangiocarcinomas, hep-
[48—52]. atocholangiocarcinomas, and liver hemangiomas can also
appear with a hypointense signal during the hepatobiliary
Gd-BOPTA phase. Therefore, detailed analysis of the features of the
nodule on other MRI sequences is still essential before mak-
Another hepatospecific contrast agent is Gd-BOPTA, or ing a diagnosis [63].
gadobenate dimeglumine (MultiHance; Bracco, Milan, Italy),
which is available in France. Hepatocyte uptake of Gd- Limitations
BOPTA is significantly lower than in comparison with
gadoxetic acid and is estimated to reach only 5%. The main Other limitations are associated with the use of gadoxetic
consequence of this relatively low uptake is that the hepato- acid. First, its efficiency is reduced in patients with major
biliary phase (also called the equilibrium phase) is delayed liver impairment. In such patients, generally those with
and observed approximately 1—2 hours after administration. Child-Pugh Class C disease, hepatocyte uptake of the con-
The use of Gd-BOPTA has been shown to help characterize trast agent can be very much reduced by the decreased
benign tumors and, in particular, distinguish between liver number of hepatocytes and their overall lack of function.
adenoma and focal nodular hyperplasia. Although Gd-BOPTA Another main limitation is the uncertain contribution of the
reportedly has high accuracy when diagnosing HCC, it has not venous or delayed phase, for which data are typically col-
been included in various international decision algorithms, lected at 3—5 minutes after injection; this is also called
which is doubtless due to the difficulties related to perform- the transition phase when using gadoxetic acid. Indeed, the
ing a second scan 1 or 2 hours later. Therefore, the rest of specificity of the hypointense signals observed during this
this chapter will exclusively discuss the use of gadoxetic phase is lower than that of the washout with an extracellu-
acid. lar contrast agent because hepatocyte uptake has already
begun. Given the late-stage washout for HCCs, the clini-
Improved characterization of cirrhotic liver cal significance of the loss of information due to the lack
nodules of a ‘‘typical’’ delayed phase remains to be assessed, even
if partially offset by additional information obtained during
When characterizing nodules in the cirrhotic liver, almost the hepatobiliary phase. Finally, as discussed in the previous
all HCCs (except for rare HCCs that continue to produce paragraph, the specificity of hypointense signals during the
the OATP8 membrane transporters, which take up gadox- hepatobiliary phase is still a problem.
etic acid) and certain H-DNs demonstrate a hypointense
signal during the hepatobiliary phase [53—55]. Thus, imag- Use in international guidelines
ing during the hepatobiliary phase improves the detection
of precancerous lesions and early-stage HCCs that still Due to these limitations and despite gadoxetic acid’s obvious
tend to display hypo- or isovascular features during the usefulness for detecting and characterizing cirrhotic liver
arterial phase and that are accordingly diagnostically chal- nodules, it took a long time for gadoxetic acid-enhanced
lenging [56—58]. However, many L-DNs and regenerative MRI to be included in the international guidelines for HCC
nodules can also appear hypointense during the hepato- management. The Japan society of hepatology was the first
biliary phase [55,59,60]. Longitudinal follow-up studies of to include gadoxetic acid in its decision algorithm in 2010
nonhypervascular nodules that appear hypointense during [64], which was later updated in 2014 [38]. In the Japanese
the hepatobiliary phase have shown that the risk that they guidelines, gadoxetic acid-enhanced MRI is the first-line
become hypervascular during the arterial phase within a diagnostic modality to be used when a nodule is detected
year following diagnosis increases with their size, and this in patients at risk for HCC. In this algorithm, hyperarterial-
risk is estimated to be 37.6% for nodules measuring > 1 cm ized nodules that demonstrate a hypointense signal during
and 77% for nodules measuring > 1.5 cm [61,62]. the hepatobiliary phase (regardless of the pattern observed
A cirrhotic liver nodule that measures > 1 cm, demon- during the portal and/or transition phase) are considered
strates hypervascularity during the arterial phase, demon- to be HCC and treated as such. Nevertheless, it should be
strates no washout during the portal or venous phase, and noted that the context in Japan differs from the Western
Diagnosis of hepatocellular carcinoma: An update on international guidelines 385

world with a greater incidence of HCC in Japan. Therefore, One of the main limitations of this technique is, again, the
from a statistical and epidemiological point of view, the ben- relative lack of specificity, and there exists a certain overlap
efit of the excellent sensitivity of gadoxetic acid-enhanced of ADC values between malignant and benign lesions. Sev-
MRI is justified, even if it means a slightly lower specificity eral studies that investigated the use of DWI to diagnose HCC
(Fig. 4). In Western countries, the hepatobiliary phase was have shown an inverse correlation between the ADC value
added to the 2014 edition of the LI-RADS classification guide- and HCC grade, with a lower ADC value noted in high-grade,
lines. A hypointense signal during the hepatobiliary phase is poorly-differentiated HCCs in comparison with low-grade,
now considered an ancillary feature indicative of HCC [42]. well-differentiated HCCs [70]. Some studies reported both
A recent study demonstrated that the use of data from the a sensitivity and a specificity of 100% for ADC in differenti-
hepatobiliary phase increased the sensitivity of the LI-RADS ating the histological grades of HCCs [71]. Typically, a high
classifications without decreasing its specificity [65]. signal intensity is observed for HCCs on high—b-value DWI
sequences, but this is not the case for dysplastic nodules
[72]. One study suggested that HCV-related HCCs more
Diffusion-weighted imaging frequently demonstrate a high signal intensity on diffusion
sequences than HBV-related HCCs [73]. An association
Diffusion-weighted imaging (DWI) has become a vital part of between an initially low ADC value, aggressiveness, and
liver imaging because of its very high sensitivity for detect- poor response to treatment (such as chemoembolization,
ing benign and malignant liver lesions [66—69]. The diffusion thermal ablation, surgical resection, or anti-angiogenic
sequence is used to calculate the apparent diffusion coeffi- therapy) was also noted, which included a higher recurrence
cient (ADC) of each voxel in the image. ADC reflects proton rate after treatment and poorer prognosis [74—77]. Recent
mobility in water and therefore indirectly provides informa- results have suggested that a 20% increase in the ADC at
tion on tissue cellularity, necrosis, vascularity, and fibrosis. 1 month after chemoembolization is indicative of good

Figure 4. Decision algorithm (2014 edition) of the Japan society of hepatology [38]. In this algorithm, gadoxetic acid-enhanced MRI is the
first-line diagnostic modality to be used when a nodule is detected in a patient at risk of developing HCC. CEUS is used as the second-line
modality when a nodule does not demonstrate hypervascularity but does show other suspicious features (such as MRI hypointensity during
the hepatobiliary phase). Hyperarterialized nodules that demonstrate a hypointense signal during the hepatobiliary phase (regardless of the
pattern observed during the portal and/or transition phase) are considered to be HCC and treated as such.
386 C. Cassinotto et al.

prognosis and low ADC at 1 month after radiofrequency of a tumor capsule, and significant tumor growth (pres-
ablation is associated with an increased risk of recurrence ence of a nodule measuring > 10 mm, 50% increase in the
[78,79]. However, the use of DWI has yet to be clearly diameter in < 6 months, or 100% increase in the diameter
defined in international guidelines. This is possibly due to over > 6 months) [42]. Nodules are graded as LR-5 if they
the lack of standardized ADC measurements, even if many combine the following characteristics:
authors have strived to provide unified criteria for its use • arterial phase enhancement, ≥ 20 mm in diameter, and
[80]. LI-RADS is the only classification system to include high ≥ 1 additional major feature, or;
DWI signal intensity (or restricted diffusion on ADC maps) as • arterial phase enhancement, ≥ 10 mm in diameter, and
an ancillary feature focused on the diagnosis of HCC [42]. ≥ 2 additional major features.

The presence of a tumor capsule — which is charac-

LI-RADS classification teristic and can be specific for progressed HCC — plays a
major role in this classification system. Absent from the
Basic principles early stages of HCC, the capsule then develops as a bar-
ricade to prevent the spread of tumor cells. The capsule
LI-RADS was established by a panel of North American then undergoes tumor cell invasion before completely dis-
experts under the auspices of the American college of appearing [81]. Therefore, the presence of a capsule is a
radiology in order to standardize the CT and MRI tech- sign of disease progression and reflects a particular step in
niques, data collection, and interpretation protocols that HCC development. The capsule plays an important role in
are used to detect HCCs or characterize nodules in tumor development, represents a complex zone where the
patients with cirrhosis or at risk of developing HCC. The tumor communicates with the adjacent parenchyma, and
objectives of this freely available online classification sys- is involved in the changes that occur in tumor drainage.
tem (http://www.acr.org/quality-safety/resources/LIRADS) The capsule is visualized as a thick, relatively regularly
[42] are to standardize the terminology used to per- shaped rim that is enhanced during the portal or delayed
form and describe imaging examinations, reduce variability phase. Care must be taken to distinguish between real cap-
in interpretation, and facilitate communication between sules and pseudocapsules that consist of peritumoral fibrotic
radiologists and treating physicians. The main advantages tissue and/or sinusoidal dilatation that can surround nod-
of this system are its regular evidence-based updates that ules in the cirrhotic liver [82,83], the latter being thinner
are founded on reports in the literature, its ordinal cate- and demonstrating less enhancement. In the same way,
gorization of lesions depending on their likelihood of being some intrahepatic cholangiocarcinomas demonstrate rim
HCC, and the diagnostic features it provides, which include enhancement during all phases and should not be confused
not only the features that describe vascularity but also take with a tumor capsule.
into account other ancillary features that favor either HCC
or a benign lesion. For the time being, its main drawbacks
are its complexity due to the great number of diagnostic fea-
Ancillary features
tures described. Regular evidence-based updates are a key
element of LI-RADS. Hence, the current criteria (2014 edi- The following ancillary features (Fig. 6) are indicative of
tion) take into account the contributions of hepatospecific HCC and can be included to increase the LI-RADS class up to
contrast agents (although, for the moment, only for ancil- LR-4 but not above:
lary features) and the LI-RADS classification adapted for • low signal intensity during the hepatobiliary phase (see
CEUS proposed in 2016 (CEUS LI-RADS 2016), even though section titled ‘‘Improved characterization of cirrhotic
AASLD removed CEUS from its decision algorithm in 2010. liver nodules’’);
Finally, the American college of radiology is preparing to • mild to moderate hyperintensity on T2-weighted images
issue two new LI-RADS classifications, one of which will focus or restricted diffusion: enhancement kinetics do not
on assessing HCC tumor response to locoregional or systemic always allow differentiation between regenerative nod-
treatment (TR LI-RADS). ules, dysplastic nodules, and early-stage HCC because all
of these lesions can sometimes present hypervascular-
ity without portal- and/or delayed-phase washout. Mild
LI-RADS classification to moderate hyperintensity with a tissue-like appearance
and/or restricted diffusion are features indicative of HCC
Major features [84];
In the 2014 edition of the LI-RADS classification system, nod- • noncapsule—like peripheral enhancement: this character-
ules and lesions are categorized as follows: LR-1, definitely istic feature of hypervascular or progressed HCC reflects
benign; LR-2, probably benign; LR-3, intermediate probabil- a typically transient enhancement of the venous drainage
ity of being an HCC or a benign lesion without the specific area in the surrounding parenchyma;
features that would allow it to be more likely characterized • mosaic architecture: this feature is frequently observed
as benign or HCC; LR-4, probably HCC; LR-5, definitely HCC; in large, progressed HCCs and reflects the existence of
LR-5 V, definitely HCC with vascular invasion; LR-M, probably discrete compartments within the lesion with different
a malignant lesion but no specific features of HCC (Fig. 5). enhancement, intensity, and fat-content patterns that
The main features indicating the presence of HCC include are generally separated by fibrous septa. This feature
enhancement during the arterial phase, tumor diameter, is not useful for diagnosing small HCCs but can help in
washout during the portal and/or venous phases, presence the differential diagnosis between HCC and intrahepatic
Diagnosis of hepatocellular carcinoma: An update on international guidelines 387

Figure 5. Schematic description of LI-RADS and its major decision features [42].

cholangiocarcinoma, which does not normally have this iron content. Indeed, iron tends to preferentially accu-
appearance; mulate in L-DNs and less frequently in H-DNs. Nodules
• ‘‘nodule-in-nodule’’ appearance: this characteristic fea- with considerable iron overload are called siderotic nod-
ture of HCC occurs when a subnodule (which usually shows ules and appear with high signal intensity on T1-weighted
hypervascularity and/or washout) is observed within images and low signal intensity on T2-weighted images.
another nodule with a different pattern. Typically, the On the contrary, most H-DNs, as well as early-stage HCC
subnodule is a progressed HCC that has developed within and progressed HCC, show lesional iron sparing;
a dysplastic nodule or early-stage HCC lesion. However, • intralesional bleeding;
this feature is only rarely observed on MRI or CT; • increase in size below the previously mentioned threshold
• fat deposition: the presence of fat within a nodule is values.
better visualized on MRI than CT. This feature character-
izes early-stage HCCs but is not highly specific because It is notable that a large majority of these ancillary fea-
fat is observed in some H-DNs as well as some L-DNs tures are observable only using MRI; this confirms the key
(although more rarely) [85]. The intranodular fat content role of this modality in diagnosing liver disease.
increases with progression from L-DN to early-stage HCC,
then decreases as the size and lesion class continue to Diagnostic performance of LI-RADS
increase. Apart from some recently reported rare forms of
classification
steatohepatitis-related HCC [86], the presence of intrale-
sional fat is generally associated with early-stage HCC The diagnostic performance of LI-RADS classification for the
and therefore, a better prognosis [87]. Naturally, it is noninvasive diagnosis of HCC depends on which category of
advisable to exclude the existence of liver adenoma or lesion is studied. As with the EASL and AASLD guidelines,
fat-infiltrated focal nodular hyperplasia; the specificity of the LR-5 category for diagnosing HCC is
• intralesional iron sparing: the signal intensity of T1- excellent (> 95%), but its sensitivity is relatively low: approx-
weighted images of dysplastic nodules depends on the imately 40% for nodules measuring < 2 cm. When the LR-4
388 C. Cassinotto et al.

Figure 6. Ancillary features of LI-RADS that are considered suggestive of either malignancy or a benign lesion [42].

and LR-5 categories are combined, the sensitivity increases Disclosure of interest
but specificity decreases [88]. Apart from hypointensity
during the hepatobiliary phase, which improves sensitivity The authors declare that they have no competing interest.
without losing in specificity [65], the contributions of the
other ancillary features to diagnostic performance have yet
to be assessed. References
[1] Binder-Foucard F, Belot A, Delafosse P, Remontet L, Woronoff
Conclusion AS, Bossard N. Cancer incidence and mortality in France over
the 1980—2012 period: solid tumors. Rev Epidemiol Sante Publ
In recent years, the techniques used with different imaging 2014;62:95—108.
modalities have been adapted to enable radiologists to dif- [2] Rahib L, Smith BD, Aizenberg R, Rosenzweig AB, Fleshman JM,
ferentiate between the numerous histological patterns seen Matrisian LM. Projecting cancer incidence and deaths to 2030:
in liver nodules and allow the detection of ever smaller HCCs the unexpected burden of thyroid, liver, and pancreas cancers
via the screening programs implemented in many countries. in the United States. Cancer Res 2014;74:2913—21.
The number and variety of radiological features have grown, [3] El-Serag HB. Hepatocellular carcinoma. N Engl J Med
2011;365:1118—27.
but they have also become more complex and require deci-
[4] Trichopoulos D, Barnia C, Lagiou P, Fedirko V, Trepo E, Jenab
sion algorithms and consensus guidelines to be regularly M, et al. Hepatocellular carcinoma and disease burden in a
updated in order to take into account new epidemiologi- European cohort: a nested casecontrol study. J Natl Cancer
cal data. In this respect, the contributions of hepatospecific Inst 2011;103:1686—95.
contrast agents is particularly promising, especially for the [5] Welzel TM, Graubard BI, Zeuzem S, El-Serag HB, Davila JA,
diagnosis of early or precancerous forms of HCC, but they McGlynn KA. Metabolic syndrome increases the risk of primary
still need to be validated in large-scale studies. CEUS will liver cancer in the United States: a study in the SEER-Medicare
also play an important role, probably by providing additional database. Hepatology 2011;54:463—71.
data to complement slice-imaging modalities. Finally, the [6] European association for the study of the liver (EASL); European
LI-RADS classification has assimilated this wealth of radio- association for the study of diabetes (EASD); European associ-
ation for the study of obesity (EASO). EASL-EASD-EASO clinical
logical features and strives to unify and standardize their
practice guidelines for the management of non-alcoholic fatty
use, even if much room is left for improvement. Faced with liver disease. J Hepatol 2016;64:1388—402.
the overwhelming wealth of radiological features, it seems [7] Cassinotto C, Charrie A, Mouries A, Lapuyade B, Hiriart JB,
likely that the use of step-by-step algorithms, such as those Vergniol J, et al. Liver and spleen elastography using supersonic
used by the Japan society of hepatology, may provide a shear imaging for the non-invasive diagnosis of cirrhosis sever-
promising solution. ity and oesophageal varices. Dig Liver Dis 2015;47:695—701.
Diagnosis of hepatocellular carcinoma: An update on international guidelines 389

[8] Ronot M, Van Beers BE, Vilgrain V. Quantification of hepatic fat: of ultrasonography in transplant patients. J Ultrasound Med
which reference should be the reference? Diagn Interv Imaging 2001;20:99—104.
2015;96:841—2. [30] Singal A, Volk ML, Waljee A, Salgia R, Higgins P, Rogers MA, et al.
[9] Cassinotto C. Elastography: French innovations in the spotlight. Meta-analysis: surveillance with ultrasound for early-stage
Diagn Interv Imaging 2016;97:1—2. hepatocellular carcinoma in patients with cirrhosis. Aliment
[10] Lafitte M, Laurent V, Soyer P, Ayav A, Balaj C, Petit I, et al. MDCT Pharmacol Ther 2009;30:37—47.
features of hepatocellular carcinoma (HCC) in non-cirrhotic [31] Sato T, Tateishi R, Yoshida H, Ohki T, Masuzaki R, Imamura J,
liver. Diagn Interv Imaging 2016;97:355—60. et al. Ultrasound surveillance for early detection of hepato-
[11] Boulin M, Delhom E, Pierredon-Foulongne MA, Cercueil JP, Guiu cellular carcinoma among patients with chronic hepatitis C.
B. Transarterial chemoembolization for hepatocellular carci- Hepatol Int 2009;3:544—50.
noma: an old method, now flavor of the day. Diagn Interv [32] Holland AE, Hecht EM, Hahn WY, Kim DC, Babb JS, Lee VS,
Imaging 2015;96:607—15. et al. Importance of small (≤ 20-mm) enhancing lesions seen
[12] Chevallier P, Baudin G, Anty R, Guibal A, Chassang M, Avril only during the hepatic arterial phase at MR imaging of the cir-
L, et al. Treatment of hepatocellular carcinomas by thermal rhotic liver: evaluation and comparison with whole explanted
ablation and hepatic transarterial chemoembolization. Diagn liver, Radiology 2005;237:938—44.
Interv Imaging 2015;96:637—46. [33] Forner A, Vilana R, Ayuso C, Bianchi L, Solé M, Ayuso JR, et al.
[13] Aubé C, Bouvier A, Lebigot J, Vervueren L, Cartier V, Oberti F. Diagnosis of hepatic nodules 20 mm or smaller in cirrhosis:
Radiological treatment of HCC: interventional radiology at the prospective validation of the noninvasive diagnostic criteria for
heart of management. Diagn Interv Imaging 2015;96:625—36. hepatocellular carcinoma. Hepatology 2008;47:97—104.
[14] Okuda K. Hepatocellular carcinoma: recent progress. Hepatol- [34] Murphy-Lavallee J, Jang HJ, Kim TK, Burns PN, Wilson SR.
ogy 1992;15:948—63. Are metastases really hypovascular in the arterial phase? The
[15] International working party. Terminology of nodular hepa- perspective based on contrast-enhanced ultrasonography. J
tocellular lesions: recommendations of the world congress Ultrasound Med 2007;26:1545—56.
of gastroenterology working group. Hepatology 1995;22: [35] Strobel D, Seitz K, Blank W, Schuler A, Dietrich CF, von
983—93. Herbay A, et al. Tumor-specific vascularization pattern of liver
[16] Kojiro M, Roskams T. Early hepatocellular carcinoma and dys- metastasis, hepatocellular carcinoma, hemangioma and focal
plastic nodules. Semin Liver Dis 2005;25:133—42. nodular hyperplasia in the differential diagnosis of 1349 liver
[17] Kudo M. Multistep human hepatocarcinogenesis: correlation of lesions in contrast-enhanced ultrasound (CEUS). Ultraschall
imaging with pathology. J Gastroenterol 2009;44:112—8. Med 2009;30:376—82.
[18] Kojiro M, Wanless IR, Alves V, Badve S, Balabaud C, Bedossa P, [36] Bhayana D, Kim TK, Jang HJ, Burns PN, Wilson SR. Hypervascu-
et al. Pathologic diagnosis of early hepatocellular carcinoma: a lar liver masses on contrast-enhanced ultrasound: the impor-
report of the international consensus group for hepatocellular tance of washout. AJR Am J Roentgenol 2010;194:977—83.
neoplasia. Hepatology 2009;49:1058. [37] Vilana R, Forner A, Bianchi L, García-Criado A, Rimola J, de
[19] Park YN. Update on precursor and early lesions of hepato- Lope CR, et al. Intrahepatic peripheral cholangiocarcinoma
cellular carcinomas. Arch Pathol Lab Med 2011;135:704—15, in cirrhosis patients may display a vascular pattern similar
http://dx.doi.org/10.1043/2010-0524-RA.1. to hepatocellular carcinoma on contrast-enhanced ultrasound.
[20] Choi JY, Lee JM, Sirlin CB. CT and MR imaging diagnosis and Hepatology 2010;51:2020—9.
staging of hepatocellular carcinoma: part I. Development, [38] Kudo M, Matsui O, Izumi N, Iijima H, Kadoya M, Imai Y, et al.
growth, and spread: key pathologic and imaging aspects. Radi- JSH Consensus-based clinical practice guidelines for the man-
ology 2014;272:635—54. agement of hepatocellular carcinoma: 2014 update by the liver
[21] Bruix J, Sherman M, Llovet JM, Beaugrand M, Lencioni R, cancer study group of Japan. Liver Cancer 2014;3:458—68.
Burroughs AK, et al. Clinical management of hepatocellular [39] Jang HJ, Kim TK, Burns PN, Wilson SR. CEUS: an essential
carcinoma. Conclusions of the Barcelona-2000 EASL confer- component in a multimodality approach to small nodules in
ence. J Hepatol 2001;35:421—30. patients at high-risk for hepatocellular carcinoma. Eur J Radiol
[22] Mancebo A, González-Diéguez ML, Cadahía V, Varela M, Pérez 2015;84:1623—35.
R, et al. Annual incidence of hepatocellular carcinoma among [40] Sporea I, Badea R, Popescu A, Spârchez Z, Sirli RL, Dănilă M,
patients with alcoholic cirrhosis and identification of risk et al. Contrast-enhanced ultrasound (CEUS) for the evaluation
groups. Clin Gastroenterol Hepatol 2013;11:95—101. of focal liver lesions-a prospective multicenter study of its use-
[23] Ascha MS, Hanouneh IA, Lopez R, Tamimi TA, Feldstein AF, fulness in clinical practice. Ultraschall Med 2014;35:259—66.
Zein NN. The incidence and risk factors of hepatocellular carci- [41] de Sio I, Iadevaia MD, Vitale LM, Niosi M, Del Prete A,
noma in patients with nonalcoholic steatohepatitis. Hepatology de Sio C, et al. Optimized contrast-enhanced ultrasonogra-
2010;51:1972—8. phy for characterization of focal liver lesions in cirrhosis: a
[24] Bruix J, Sherman M, Practice guidelines committee, Ameri- single-center retrospective study. United Eur Gastroenterol J
can association for the study of liver diseases. Management 2014;2:279—87.
of hepatocellular carcinoma. Hepatology 2005;42:1208—36. [42] American college of radiology, liver imaging reporting and
[25] Bruix J, Sherman M. Management of hepatocellular carcinoma: data system. http://www.acr.org/quality-safety/resources/
an update. Hepatology 2011;53:1020—2. LIRADS, 2016.(accessed February 3, 2017).
[26] European association for the study of the liver, European orga- [43] Vilgrain V, Esvan M, Ronot M, Caumont-Prim A, Aubé C, Chatel-
nisation for research and treatment of cancer. EASL-EORTC lier G. A meta-analysis of diffusion-weighted and gadoxetic
clinical practice guidelines: management of hepatocellular acid-enhanced MR imaging for the detection of liver metas-
carcinoma. J Hepatol 2012;56:908—43. tases. Eur Radiol 2016;26:4595—615.
[27] The French METAVIR cooperative study group. Intraobserver [44] Raman SS, Leary C, Bluemke DA, Amendola M, Sahani D,
and interobserver variations in liver biopsy interpretation in McTavish JD, et al. Improved characterization of focal liver
patients with chronic hepatitis C. Hepatology 1994;20:15—20. lesions with liver-specific gadoxetic acid disodium-enhanced
[28] Bolondi L. Screening for hepatocellular carcinoma in cirrhosis. magnetic resonance imaging: a multicenter phase 3 clinical
J Hepatol 2003;39:1076—84. trial. J Comput Assist Tomogr 2010;34:163—72.
[29] Kim CK, Lim JH, Lee WJ. Detection of hepatocellular car- [45] Koh DM, Collins DJ, Wallace T, Chau I, Riddell AM. Combin-
cinomas and dysplastic nodules in cirrhotic liver: accuracy ing diffusion-weighted MRI with Gd-EOB-DTPA-enhanced MRI
390 C. Cassinotto et al.

improves the detection of colorectal liver metastases. Br J [61] Kumada T, Toyoda H, Tada T, Sone Y, Fujimori M, Ogawa S, et al.
Radiol 2012;85:980—9. Evolution of hypointense hepatocellular nodules observed only
[46] Katsube T, Okada M, Kumano S, Hori M, Imaoka I, Ishii K, in the hepatobiliary phase of gadoxetate disodium-enhanced
et al. Estimation of liver function using T1 mapping on Gd-EOB- MRI. AJR Am J Roentgenol 2011;197:58—63.
DTPA-enhanced magnetic resonance imaging. Invest Radiol [62] Sano K, Ichikawa T, Motosugi U, Ichikawa S, Morisaka H,
2011;464:277—83. Enomoto N, et al. Outcome of hypovascular hepatic nodules
[47] Yamada A, Hara T, Li F, Fujinaga Y, Ueda K, Kadoya M, with positive uptake of gadoxetic acid in patients with cirrho-
et al. Quantitative evaluation of liver function with use sis. Eur Radiol 2017;27:518—25.
of gadoxetate disodium-enhanced MR imaging. Radiology [63] Chong YS, Kim YK, Lee MW, Kim SH, Lee WJ, Rhim HC, et al.
2011;260:727—33. Differentiating mass-forming intrahepatic cholangiocarcinoma
[48] Ichikawa T, Saito K, Yoshioka N, Tanimoto A, Gokan T, Takehara from atypic al hepatocellular carcinoma using gadoxetic acid-
Y, et al. Detection and characterization of focal liver lesions: enhanced MRI. Clin Radiol 2012;67:766—73.
a Japanese phase III, multicenter comparison between gadox- [64] Kudo M, Izumi N, Kokudo N, Matsui O, Sakamoto M, Nakashima
etic acid disodium-enhanced magnetic resonance imaging and O, et al. Management of hepatocellular carcinoma in Japan:
contrast-enhanced computed tomography predominantly in consensus-based clinical practice guidelines proposed by the
patients with hepatocellular carcinoma and chronic liver dis- Japan society of hepatology (JSH) 2010 updated version. Dig
ease. Invest Radiol 2010;45:133—41. Dis 2011;29:339—64.
[49] Merkle EM, Zech CJ, Bartolozzi C, Bashir MR, Ba-Ssalamah A, [65] Chen N, Motosugi U, Morisaka H, Ichikawa S, Sano K,
Huppertz A, et al. Consensus report from the 7th Interna- Ichikawa T, et al. Added value of a gadoxetic acid-enhanced
tional forum for liver magnetic resonance imaging. Eur Radiol hepatocyte-phase image to the LI-RADS system for diagnos-
2016;26:674—82. ing hepatocellular carcinoma. Magn Reson Med Sci 2016;15:
[50] Sirlin CB, Hussain HK, Jonas E, Kanematsu M, Min Lee J, 49—59.
Merkle EM, et al. Consensus report from the 6th International [66] Park MS, Kim S, Patel J, Hajdu CH, Do RK, Mannelli L, et al.
forum for liver MRI using gadoxetic acid. J Magn Reson Imaging Hepatocellular carcinoma: detection with diffusion-weighted
2014;40:516—29. versus contrast-enhanced magnetic resonance imaging in pre-
[51] Jhaveri K, Cleary S, Audet P, Balaa F, Bhayana D, Burak K, transplant patients. Hepatology 2012;56:140—8.
et al. Consensus statements from a multidisciplinary expert [67] Soyer P, Boudiaf M, Place V, Sirol M, Pautrat K, Vignaud A,
panel on the utilization and application of a liver-specific et al. Preoperative detection of hepatic metastases: com-
MRI contrast agent (gadoxetic acid). AJR Am J Roentgenol parison of diffusion-weighted. T2-weighted fast spin echo
2015;204:498—509. and gadolinium-enhanced MR imaging using surgical and
[52] Choi SH, Byun JH, Lim YS, Yu E, Lee SJ, Kim SY, et al. Diagnostic histopathologic findings as standard of reference. Eur J Radiol
criteria for hepatocellular carcinoma ≤ 3 cm with hepatocyte- 2011;80:245—52.
specific contrast-enhanced magnetic resonance imaging. J [68] Dohan A, Soyer P, Guerrache Y, Hoeffel C, Gavini JP, Kaci R,
Hepatol 2016;64:1099—107. et al. Focal nodular hyperplasia of the liver: diffusion-weighted
[53] Bartolozzi C, Battaglia V, Bargellini I, et al. Contrast-enhanced magnetic resonance imaging characteristics using high B val-
magnetic resonance imaging of 102 nodules in cirrhosis: cor- ues. J Comput Assist Tomogr 2014;38:96—104.
relation with histological findings on explanted livers. Abdom [69] Cassinotto C, Feldis M, Vergniol J, Mouries A, Cochet H,
Imaging 2013;38:290—6. Lapuyade B, et al. MR relaxometry in chronic liver diseases:
[54] Golfieri R, Renzulli M, Lucidi V, Corcioni B, Trevisani F, Bolondi comparison of T1 mapping. T2 mapping, and diffusion-
L. Contribution of the hepatobiliary phase of Gd-EOB-DTPA- weighted imaging for assessing cirrhosis diagnosis and severity.
enhanced MRI to dynamic MRI in the detection of hypo-vascular Eur J Radiol 2015;84:1459—65.
small (< 2 cm) HCC in cirrhosis. Eur Radiol 2011;21:1233—42. [70] Li X, Zhang K, Shi Y, Wang F, Meng X. Correlations between
[55] Leoni S, Piscaglia F, Golfieri R, Camaggi V, Vidili G, Pini P, the minimum and mean apparent diffusion coefficient values
et al. The impact of vascular and nonvascular findings on of hepatocellular carcinoma and tumor grade. J Magn Reson
the noninvasive diagnosis of small hepatocellular carcinoma Imaging 2016;44:1442—7.
based on the EASL and AASLD criteria. Am J Gastroenterol [71] Granata V, Fusco R, Catalano O, Guarino B, Granata F, Tatangelo
2010;105:599—609. F, et al. Intravoxel incoherent motion (IVIM) in diffusion-
[56] Bolondi L, Gaiani S, Celli N, Golfieri R, Grigioni WF, Leoni weighted imaging (DWI) for hepatocellular carcinoma: corre-
S, et al. Characterization of small nodules in cirrhosis by lation with histologic grade. Oncotarget 2016;7:79357—64.
assessment of vascularity: the problem of hypovascular hepa- [72] Xu PJ, Yan FH, Wang JH, Shan Y, Ji Y, Chen CZ. Contribution of
tocellular carcinoma. Hepatology 2005;42:27—34. diffusion-weighted magnetic resonance imaging in the charac-
[57] Choi JY, Lee JM, Sirlin CB. CT and MR imaging diagnosis terization of hepatocellular carcinomas and dysplastic nodules
and staging of hepatocellular carcinoma: part II. Extracellular in cirrhotic liver. J Comput Assist Tomogr 2010;34:506—12.
agents, hepatobiliary agents, and ancillary imaging features. [73] Dunst D, Ream JM, Khalef V, Hajdu CH, Rosenkrantz AB. Com-
Radiology 2014;273:30—50. parison of MRI features of pathologically proven hepatocellular
[58] Kogita S, Imai Y, Okada M, et al. Gd-EOB-DTPA-enhanced carcinoma between patients with hepatitis B and hepatitis C
magnetic resonance images of hepatocellular carcinoma: cor- infection. Clin Imaging 2016;40:352—6.
relation with histological grading and portal blood flow. Eur [74] Kokabi N, Ludwig JM, Camacho JC, Xing M, Mittal PK, Kim HS.
Radiol 2010;20:2405—13. Baseline and early MR apparent diffusion coefficient quantifica-
[59] Motosugi U, Ichikawa T, Sano K, Sou H, Onohara K, Muhi A, tion as a predictor of response of unresectable hepatocellular
et al. Outcome of hypovascular hepatic nodules revealing no carcinoma to doxorubicin drug-eluting bead chemoemboliza-
gadoxetic acid uptake in patients with chronic liver disease. J tion. J Vasc Interv Radiol 2015;26:1777—86.
Magn Reson Imaging 2011;34:88—94. [75] Nakanishi M, Chuma M, Hige S, Omatsu T, Yokoo H, Nakanishi K,
[60] Kwon HJ, Byun JH, Kim JY, Hong GS, Won HJ, Shin YM, et al. et al. Relationship between diffusion-weighted magnetic reso-
Differentiation of small (≤ 2 cm) hepatocellular carcinomas nance imaging and histological tumor grading of hepatocellular
from small benign nodules in cirrhotic liver on gadoxetic acid- carcinoma. Ann Surg Oncol 2012;19:1302—9.
enhanced and diffusion-weighted magnetic resonance images. [76] Nishie A, Tajima T, Asayama Y, Ishigami K, Kakihara D,
Abdom Imaging 2015;40:64—75. Nakayama T, et al. Diagnostic performance of apparent
Diagnosis of hepatocellular carcinoma: An update on international guidelines 391

diffusion coefficient for predicting histological grade of hepa- [82] Ng IO, Lai EC, Ng MM, Fan ST. Tumor encapsulation in hepa-
tocellular carcinoma. Eur J Radiol 2011;80:e29—33. tocellular carcinoma: a pathologic study of 189 cases. Cancer
[77] Yuan Z, Zhang J, Yang H, Ye XD, Xu LC, Li WT. Diffusion- 1992;70:45—9.
weighted MR imaging of hepatocellular carcinoma: current [83] Ishigami K, Yoshimitsu K, Nishihara Y, Irie H, Asayama Y, Tajima
value in clinical evaluation of tumor response to locoregional T, et al. Hepatocellular carcinoma with a pseudocapsule on
treatment. J Vasc Interv Radiol 2016;27:20—30. gadolinium-enhanced MR images: correlation with histopatho-
[78] Kokabi N, Camacho JC, Xing M, Edalat F, Mittal PK, logic findings. Radiology 2009;250:435—43.
Kim HS. Immediate post-doxorubicin drug-eluting beads [84] Piana G, Trinquart L, Meskine N, Barrau V, Beers BV, Vilgrain
chemoembolization MR apparent diffusion coefficient quan- V, et al. imaging criteria with a diffusion-weighted sequence
tification predicts response in unresectable hepatocellular for the diagnosis of hepatocellular carcinoma in chronic liver
carcinoma: a pilot study. J Magn Reson Imaging 2015;42: diseases. J Hepatol 2011;55:126—32.
981—9. [85] Kanematsu M, Kondo H, Goshima S, Tsuge Y, Watanabe H.
[79] Barat M, Fohlen A, Cassinotto C, Jannot AS, Dautry R, Magnetic resonance imaging of hepatocellular carcinoma.
Pelage JP, et al. One-month apparent diffusion coeffi- Oncology 2008;75:65—71.
cient correlates with response to radiofrequency ablation [86] Salomao M, Yu WM, Brown Jr RS, Emond JC, Lefkowitch JH.
of hepatocellular carcinoma. J Magn Reson Imaging 2016, Steatohepatitic hepatocellular carcinoma (SH-HCC): a distinc-
http://dx.doi.org/10.1002/jmri.25521. tive histological variant of HCC in hepatitis C virus-related
[80] Taouli B, Beer AJ, Chenevert T, Collins D, Lehman C, Matos C, cirrhosis with associated NAFLD/NASH. Am J Surg Pathol
et al. Diffusion-weighted imaging outside the brain: consensus 2010;34:1630—6.
statement from an ISMRM-sponsored workshop. J Magn Reson [87] Siripongsakun S, Lee JK, Raman SS, Tong MJ, Sayre J, Lu DS.
Imaging 2016;44:521—40. MRI detection of intratumoral fat in hepatocellular carcinoma:
[81] Iguchi T, Aishima S, Sanefuji K, Fujita N, Sugimachi K, potential biomarker for a more favorable prognosis. AJR Am J
Gion T, et al. Both fibrous capsule formation and extracap- Roentgenol 2012;199:1018—25.
sular penetration are powerful predictors of poor survival [88] Darnell A, Forner A, Rimola J, Reig M, García-Criado Á, Ayuso C,
in human hepatocellular carcinoma: a histological assess- et al. Liver imaging reporting and data system with MR imaging:
ment of 365 patients in Japan. Ann Surg Oncol 2009;16: evaluation in nodules 20 mm or smaller detected in cirrhosis at
2539—46. screening US. Radiology 2015;275:698—707.