These authors contributed equally to this manuscript.
Article first published online: 11 MAY 2009
Copyright © 2009 American Association for the Study of Liver Diseases
Volume 50, Issue 3, pages 791–798, September 2009
How to Cite
Rimola, J., Forner, A., Reig, M., Vilana, R., de Lope, C. R., Ayuso, C. and Bruix, J. (2009), Cholangiocarcinoma in cirrhosis: Absence of contrast washout in delayed phases by magnetic resonance imaging avoids misdiagnosis of hepatocellular carcinoma. Hepatology, 50: 791–798. doi: 10.1002/hep.23071
This study has been performed within the cooperation framework established by the Transversal Cancer Action approved by the Council of Ministers on October 11, 2007 in accordance with the agreement between The Carlos III Health Institute (ISCIII), which is an autonomous entity currently belonging to the Ministry of Science and Innovation, and the Spanish Biomedical Research Network (CIBER) for the area of hepatic and digestive disorders.
Potential conflict of interest: Nothing to report.
- Issue published online: 27 AUG 2009
- Article first published online: 11 MAY 2009
- Accepted manuscript online: 11 MAY 2009 12:00AM EST
- Manuscript Accepted: 5 MAY 2009
- Manuscript Received: 17 MAR 2009
- Instituto de Salud Carlos III. Grant Numbers: PI 06/132, PI 08/0146, PI 05/645
- BBVA foundation
This study assesses the magnetic resonance (MR) features of intrahepatic cholangiocarcinoma (ICC) in patients with cirrhosis with specific analysis of the contrast enhancement pattern. Cholangiocarcinoma may show increased contrast uptake in the arterial phase, and, if washout in the delayed venous phase were to be detected, the noninvasive diagnostic criteria proposed in the American Association for the Study of Liver Diseases guidelines would be refuted. We reviewed the MR findings of 25 patients with cirrhosis with 31 histologically confirmed ICC nodules. Signal intensity on basal T1-weighted and T2-weighted images and characteristics of enhancement after contrast administration on arterial, portal, and delayed phase were registered. Enhancement pattern was defined according to the behavior of the lesions in each phase, and dynamic pattern was described according to the progression of enhancement throughout the different phases. The most frequent pattern displayed by ICC was a progressive contrast uptake (80.6%). Stable contrast enhancement was registered in 19.4%. None of the ICCs showed a washout pattern, a profile that is specific for hepatocellular carcinoma (HCC). The ICC dynamic behavior differed significantly according to tumor size: progressive enhancement pattern was the most frequent (20 of 25 cases) in lesions larger than 20 mm, whereas the stable pattern was mainly identified in nodules smaller than 20 mm. The most characteristic MR contrast pattern in ICC in cirrhosis is a progressive contrast uptake throughout the different phases, whereas contrast washout at delayed phases is not observed. Because stable enhancement pattern without washout also can be registered in small HCC nodules, the evaluation of delayed phase is mandatory for a proper nodule characterization. If washout is not registered, a biopsy should be mandatory for diagnosis. (HEPATOLOGY 2009.)
The incidence of hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (ICC) has experienced a marked increase in recent decades.1, 2 HCC appears mainly in the setting of cirrhosis, and this neoplasia has become the main cause of death in this population.3 In addition, cirrhosis, mainly secondary to chronic infection with hepatitis C virus, has been recognized as an important risk factor for ICC development, and this association may be the cause of the increasing incidence of ICC in recent years.4, 5 It is critical to accurately distinguish both entities, because the prognosis and treatment are markedly different. In fact, patients diagnosed with liver cirrhosis are the target of early detection plans to diagnose HCC at an early stage, when it may be treated by effective options such as resection, transplantation, or percutaneous ablation.6, 7 Screening is done by imaging techniques, namely, ultrasound, and on detection of a suspicious nodule, the diagnosis may be established by biopsy or by imaging techniques.6 These have to be highly specific to avoid false-positive results. Major information is available to characterize the HCC dynamic imaging pattern after contrast administration at ultrasound, computed tomography, and magnetic resonance: intense arterial uptake followed by washout in the venous phase by dynamic imaging techniques. This has allowed us to define noninvasive imaging criteria to diagnose HCC in the setting of liver cirrhosis.6, 8 However, as previously commented, HCC is not the sole malignancy that may appear in liver cirrhosis. ICC also may complicate cirrhosis, and in most centers it constitutes a contraindication for liver transplantation.4, 9 Several reports have described the imaging appearance of ICC by magnetic resonance (MR) in persons without cirrhosis10–13 and suggest that in large tumors within a noncirrhotic liver the diagnostic challenge may not be relevant. Contrarily, on detection of a small nodule during screening in a patient with liver cirrhosis, the diagnostic difficulties may be higher. No previous studies have specifically reviewed the ICC MR findings in patients with cirrhosis, a specific population in whom portal hypertension and associated circulatory disturbances may induce a particular profile. Both HCC and ICC can have an increased arterial vascularization, and there is no information about the potential relationship between imaging profile and tumor size, and to which extent small arterialized ICCs may present washout in the venous/delayed phase. If this should occur, the American Association for the Study of Liver Diseases diagnostic criteria would be invalidated. Hence, data collected in series with advanced ICC will not serve to define the radiology pattern at early stages.14
The purpose of this study was to assess the MR features of ICC in the setting of cirrhosis, taking into account the nodule size and placing special emphasis on the enhancement pattern and the differential diagnosis with HCC.
Patients and Methods
We retrospectively reviewed the radiological images of all patients with ICC and cirrhosis consecutively registered in our institution between September 2001 and December 2008. The inclusion criteria were as follows: (1) pathology-confirmed diagnosis of ICC by percutaneous biopsy or surgical specimen analysis. Patients with mixed hepatocellular–cholangiocarcinoma were excluded; (2) patients with an abdominal MR performed at our institution following our standard protocol for focal hepatic lesion (see below); (3) patients with cirrhosis diagnosed by imaging (nodular or irregular liver surface and liver parenchyma heterogenicity with evidence of portal hypertension),15–18 pathology, or clinical criteria.
MR was performed in all patients with a 1.5-T MR system (Symphony, Siemens Medical Systems, Erlangen, Germany; and SIGNA CVi, General Electric Medical System, Milwaukee, WI), using a phased-array coil for signal detection. All patients underwent transverse T1-weighted in-phase and out-of-phase gradient echo, slice thickness of 5 mm, transverse T2-weighted breath-hold, slice thickness of 5 mm, and axial dynamic multiphasic contrast-enhanced three-dimensional image of the liver with fat suppression before and after gadolinium contrast administration (gadodiamide 0.5 mmol/L Ominscan-Amersham, Madrid, Spain), with a slice thickness of 2.5 to 3 mm. Arterial-phase images were acquired approximately 20 seconds after contrast injection. Portal venous and delayed-phase images were acquired 60-65 and 100-110 seconds thereafter. Finally, a T1-weighted gradient echo with fat suppression MR imaging was performed 5 minutes after contrast injection.
MR findings were retrospectively analyzed in consensus by two abdominal radiologists (J.R. and C.A.) with 5 and 20 years of experience in abdominal radiology, respectively. Each lesion was evaluated as following: size of the lesion, hepatic capsule retraction, and presence of satellite nodules and central scar.
The signal on pre-contrast T1-weighted and T2-weighted sequences was recorded. After endovenous contrast administration, the signal through each of the different phases was registered as follows: (1) globally hyperintense: increased signal, relative to the liver parenchyma, involving the totality of the lesion; (2) partially hyperintense: increased signal involving more than 25% of the lesion, except the central area, (3) peripherally hyperintense: increased signal limited to the periphery of the lesion, involving less than 25% of its area, resembling a rim-like pattern (Fig. 1).
Additionally, a dynamic pattern of enhancement was defined according to the analysis of the progression of endovenous contrast enhancement over the progressive different phases of the study, as follows: (1) stable, persistent contrast enhancement: the nodule enhancement remains invariable through the arterial and venous phases; (2) progressive contrast enhancement: the nodule enhances progressively over time, reaching maximal intensity in delayed phases; (3) washout: intense contrast uptake during the arterial phase followed by contrast washout in delayed phases (Figs. 2-4)
Data were analyzed on a by-lesion basis. Quantitative variables were expressed as median and range and categorically as count and proportions. Differences in signal intensity in baseline and postcontrast sequences and in dynamic enhancement pattern according to nodule size (≤2 cm, 2-3 cm, and >3 cm) were evaluated by the chi-squared test/Fisher's exact test for categorical variables. A conventional P value less than 0.05 was considered significant. Calculations were done with SPSS package version 13 (SPSS Inc., Chicago, IL).
Baseline Patient Characteristics.
A total of 25 patients (17 men and 8 women; median age, 62; range, 49-79 years) met all the criteria to collect the ICC cohort. Twenty patients had uninodular ICC (80%); and five, multifocal (four patients had two nodules [16%], and one patient had three nodules [4%]). In the five cases with more than one nodule, all were located in a different segment and thus were not read as a peripheral satellite nodule.
Cirrhosis was confirmed by pathology assessment in 13 patients and by unequivocal clinical and imaging criteria in the remaining cases. Causes of liver cirrhosis included chronic hepatitis C virus infection (14 patients), alcohol abuse (six patients), hemochromatosis (one patient), chronic hepatitis C virus infection combined with alcohol abuse (two patients), and cryptogenetic (two patients). Child-Pugh class stage19 at diagnosis was A (24 patients) and B (one patient). Final diagnosis of ICC was confirmed by resected specimen analysis in 13 patients and by ultrasonography-guided percutaneous liver biopsy in 12 patients. The main patient characteristics are summarized in Table 1.
|Age median (range), (years)||62 (49–79)|
|HCV + ethanol||2|
|Child-Pugh A/B (n)||24/1|
|AST median (range), (UI/L)||48 (24–149)|
|ALT median (range), (UI/L)||50 (15–167)|
|AP median (range), (UI/L)||197 (121–846)|
|GGT median (range), (UI/L)||65 (17–537)|
|Prothrombin ratio median (range), (%)||90 (60–100)|
|Bilirubin median (range), (mg/dL)||1 (0.5–17.5)|
|Platelets median (range), (109/L)||152 (43–262)|
|Albumin median (range), (g/dL)||42 (27–47)|
|Nodule size median (range), (mm)||25 (13–87)|
|≤20 mm (n)||9|
|21–30 mm (n)||11|
|>20 mm (n)||11|
|Number of nodules: 1/2/3 (n)||20/4/1|
Morphologic MR Features.
A total of 31 nodules with histological diagnosis of ICC were analyzed, and the main characteristics are summarized in Tables 2 and 3. The median size was 25 mm (range, 13-87). In 9 of 31 cases (29%), the nodule was equal to or smaller than 2 cm; in 11 cases (35.5%), between 2 and 3 cm; and in the remaining 11 nodules (35.5%), the nodule size was larger than 3 cm. A tumor scar was present in seven nodules (22.6%), retraction of the hepatic capsule was seen in seven cases (22.6%), and satellites were demonstrated in eight cases (25.8%).
|Id||Age||Size||T1-w||T2-w||Scar||Satellites||Capsular Retraction||Arterial Phase Enhancement||Portal Phase Enhancement||Delayed Phase Enhancement||Dynamic Pattern|
|1||74||19||Hypo||Iso||No||No||No||Globally hyper||Globally hyper||Globally hyper||Stable|
|2||64||28||Hypo||Iso||Yes||Yes||No||Peripherally hyper||Partially hyper||Partially hyper||Progressive|
|3||58||70||Hypo||Hyper||No||Yes||Yes||Peripherally hyper||Partially hyper||Partially hyper||Progressive|
|4||79||45||Hypo||Heterog||Yes||No||Yes||Partially hyper||Partially hyper||Partially hyper||Progressive|
|5||63||67||Hypo||Hyper||Yes||Yes||Yes||Peripherally hyper||Partially hyper||Partially hyper||Progressive|
|6||54||87||Hypo||Hyper||No||No||No||Partially hyper||Partially hyper||Partially hyper||Progressive|
|7||71||44||Hypo||Hyper||No||Yes||Yes||Peripherally hyper||Partially hyper||Partially hyper||Progressive|
|8||61||30||Iso||Hyper||No||No||No||Peripherally hyper||Partially hyper||Globally hyper||Progressive|
|9||64||25||Hypo||Hyper||No||No||No||Peripherally hyper||Partially hyper||Globally hyper||Progressive|
|14||Hypo||Hyper||No||Yes||Yes||Peripherally hyper||Peripherally hyper||Globally hyper||Progressive|
|10||69||51||Hypo||Heterog||Yes||No||No||Peripherally hyper||Partially hyper||Partially hyper||Progressive|
|20||Hypo||Hyper||No||Yes||No||Peripherally hyper||Peripherally Hyper||Peripherally hyper||Stable|
|15||Hypo||Hyper||No||No||No||Peripherally hyper||Peripherally hyper||Peripherally hyper||Stable|
|11||49||24||Hypo||Hyper||Yes||No||Yes||Peripherally hyper||Peripherally hyper||Peripherally hyper||Stable|
|12||63||24||Hypo||Heterog||No||No||No||Peripherally hyper||Peripherally hyper||Globally hyper||Progressive|
|16||Hypo||Hyper||No||Yes||No||Peripherally hyper||Partially hyper||Globally hyper||Progressive|
|13||53||36||Hypo||Hyper||Yes||No||No||Peripherally hyper||Partially hyper||Partially hyper||Progressive|
|14||77||23||Iso||Hyper||No||Yes||No||Peripherally hyper||Partially hyper||Globally hyper||Progressive|
|15||56||15||Hypo||Hyper||No||No||No||Peripherally hyper||Partially hyper||Globally hyper||Progressive|
|16||62||13||Iso||Iso||No||No||No||Globally hyper||Globally hyper||Globally hyper||Stable|
|20||Iso||Hyper||No||No||No||Peripherally hyper||Partially hyper||Globally hyper||Progressive|
|17||59||29||Hypo||Hyper||No||No||Yes||Peripherally hyper||Partially hyper||Globally hyper||Progressive|
|18||49||25||Hypo||Hypo||Yes||No||No||Peripherally hyper||Peripherally hyper||Partially hyper||Progressive|
|19||51||28||Hypo||Hyper||No||No||No||Peripherally hyper||Peripherally hyper||Partially hyper||Progressive|
|20||57||22||Hypo||Hyper||No||No||No||Peripherally hyper||Peripherally hyper||Globally hyper||Progressive|
|21||78||34||Hypo||Hyper||No||No||No||Peripherally hyper||Partially hyper||Globally hyper||Progressive|
|22||57||24||Hypo||Hyper||No||No||No||Partially hyper||Partially hyper||Globally hyper||Progressive|
|23||61||35||Iso||Iso||No||No||No||Peripherally hyper||Partially hyper||Partially hyper||Progressive|
|20||Iso||Iso||No||No||No||Peripherally hyper||Peripherally hyper||Partially hyper||Progressive|
|24||68||38||Hypo||Hyper||No||No||No||Peripherally hyper||Partially hyper||Partially hyper||Progressive|
|25||66||35||Hypo||Hyper||No||No||No||Peripherally hyper||Peripherally hyper||Peripherally hyper||Stable|
|Total (n = 31)||< 20 mm (n = 9)||21–30 mm (n = 11)||>30 mm (n = 11)||P|
|Hypointense, n (%)||25 (80.6%)||6 (66.7%)||9 (81.8%)||10 (90.9%)||NS|
|Isointense, n (%)||6 (19.4%)||3 (33.3%)||2 (18.2%)||1 (9.1%)|
|Hyperintense, n (%)||22 (71%)||6 (66.7%)||8 (72.7%)||8 (72.7%)||NS|
|Isointense, n (%)||5 (16.1%)||3 (30.3%)||1 (9.1%)||1 (9.1%)|
|Heterogeneous, n (%)||3 (9.7%)||0 (0%)||1 (9.1%)||2 (18.2%)|
|Hypointense, n (%)||1 (3.2 %)||0 (0%)||1 (9.1%)||0 (0%)|
|Yes, n (%)||8 (25.8%)||3 (33.3%)||2 (18.2%)||3 (27.3%)||NS|
|No, n (%)||23 (74.2%)||6 (66.7%)||9 (81.8%)||8 (72.7%)|
|Yes, n (%)||7 (22.6%)||0 (0%)||3 (27.3%)||4 (36.4%)||NS|
|No, n (%)||24 (77.4%)||9 (100%)||8 (72.7%)||7 (63.6%)|
|Yes, n (%)||7 (22.6%)||1 (11.1%)||2 (18.2%)||4 (36.4%)||NS|
|No, n (%)||24 (77.4%)||8 (88.9%)||9 (81.8%)||7 (63.6%)|
|Signal in arterial phase|
|Globally hyperintense, n (%)||2 (6.5%)||2 (22.2%)||0 (0%)||0 (0%)||NS|
|Partially hyperintense, n (%)||3 (9.7%)||0 (0%)||1 (9.1%)||2 (18.2%)|
|Peripherally hyperintense, n (%)||26 (83.8%)||7 (77.8%)||10 (90.9%)||9 (81.8%)|
|Signal in portal phase|
|Globally hyperintense, n (%)||2 (6.5%)||2 (22.2%)||0 (0%)||0 (0%)||0.03|
|Partially hyperintense, n (%)||19 (61.3%)||3 (33.3%)||6 (54.5%)||10 (90.9%)|
|Peripherally hyperintense, n (%)||10 (32.2%)||4 (44.5%)||5 (45.4%)||1 (9.1%)|
|Signal in delayed phase|
|Globally hyperintense, n (%)||14 (45.2%)||6 (66.7%)||7 (63.6%)||1 (9.1%)||0.014|
|Partially hyperintense, n (%)||13 (41.9%)||1 (11.1%)||3 (27.3%)||9 (81.8%)|
|Peripherally hyperintense, n (%)||4 (12.9%)||2 (22.2%)||1 (9.1%)||1 (9.1%)|
|Pattern of enhancement|
|Progressive, n (%)||25 (80.6%)||5 (55.6%)||10 (90.9%)||10 (90.9%)||0.04*|
|Stable, n (%)||6 (19.4%)||4 (44.4%)||1 (9.1%)||1 (9.1%)|
|Washout, n (%)||0 (0%)||0 (0%)||0 (0%)||0 (0%)|
On T1-weighted sequences, 25 nodules (80.6 %) appeared as hypointense and the remaining six lesions (19.4%) as isointense. On T2-weighted images, 22 nodules (71%) appeared as hyperintense, three (9.7%) as heterogeneous mass with areas of different signal, five (16.1%) as isointense relative to the hepatic parenchyma, and one (3.2%) as hypointense. On arterial phase, 26 lesions (83.8%) were peripherally hyperintense, three nodules (9.7%) partially hyperintense, and two lesions (6.5%) were globally hyperintense. On portal venous phase, 19 lesions (61.3%) were partially hyperintense, 10 lesions (32.3%) peripherally hyperintense, and two lesions (6.5%) were globally hyperintense. Finally, on delayed phase, 14 nodules (45.2%) appeared globally hyperintense; in 13 lesions (41.9%) the signal was partially hyperintense; and the remaining four nodules (12.9%) displayed a peripheral hyperintensity. None of the ICC nodules was isointense or hypointense compared with the surrounding liver parenchyma during the dynamic study.
Patterns of Enhancement.
Analysis of the vascular dynamic enhancement pattern throughout the different phases of the study showed that 25 lesions (80.6%) had a progressive contrast enhancement: in 12 nodules, the contrast uptake progressed until complete enhancement of the lesion, and in 13 lesions the contrast uptake progressed but a small central area remained unenhanced. The remaining six lesions (19.4%) showed a stable contrast enhancement during the dynamic study: four lesions displayed a stable, peripheral, rim-like contrast enhancement, and two lesions showed a global contrast enhancement during the dynamic study (Table 4).
|Pattern of Enhancement of ICC||Total (n = 31)||≤ 20 mm (n = 9)||21–30 mm (n = 11)||> 30 mm (n = 11)||P|
|Progressive, n (%)||25 (80.6%)||5 (55.6%)||10 (90.9%)||10 (90.9%)|
|Complete, n (%)||12 (38.7%)||4 (44.4%)||7 (65.5%)||1 (9.1%)||0.01|
|Incomplete, n (%)||13 (41.9%)||1 (11.1%)||3 (27.3%)||9 (81.8%)|
|Stable, n (%)||6 (19.4%)||4 (44.4%)||1 (9.1%)||1 (9.1%)|
|Peripheral, rim-like, n (%)||4 (12.9%)||2 (22.2%)||1 (9.1%)||1 (9.1%)|
|Complete, n (%)||2 (6.5%)||2 (22.2%)||0 (0%)||0 (0%)|
MR Findings According the Nodule Size.
There were no significant differences in MR appearance of ICC on T1-weighted and T2-weighted sequences, presence of satellites, capsular retraction, or intralesional scar and on the arterial phase enhancement according to nodule size (≤2, 2-3, and >3 cm) (Table 3). However, the only two nodules globally hyperintense in arterial phase were smaller than 2 cm. In the portal phase, no nodule larger than 2 cm displayed global hyperintense signal, and in almost all nodules greater than 3 cm (10 cases, 90.9%), the nodule signal was partially hyperintense (P = 0.03). In the delayed phase, 13 of 20 nodules smaller than 3 cm (65%) showed a global enhancement compared with only one of 11 nodules (9.1%) larger than 3 cm (P = 0.014).
Taking into account the dynamic vascular pattern previously defined, 20 of 22 nodules (90.9%) larger than 2 cm displayed a progressive contrast enhancement compared with only five of nine (55.6%) nodules smaller than 2 cm. Contrarily, the stable pattern was more frequently identified in nodules smaller than 2 cm (four of nine nodules, 44.4%) than in larger nodules (2 of 22, 9.1%) (P = 0.043). Finally, the washout pattern was not identified in any patient. These differences related to size achieved statistical significance (P = 0.001).
Our findings provide valuable information for the accurate diagnosis of small hepatic nodules detected in patients with cirrhosis. Screening of this population is advocated to detect HCC at an early stage. However, not all nodules arising within a cirrhotic liver are malignant, nor are they always an HCC. It is well known that cirrhosis is a risk factor for ICC, and hence, this entity should be taken into consideration because the clinical management is sharply different, and thus, the need to distinguish between the two is not trivial.4 Liver transplantation is an established treatment option for HCC, but it is not usually considered for ICC. In fact, in the current United Network for Organ Sharing criteria for liver allocation, clear-cut diagnostic criteria for HCC are demanded to be enlisted for transplantation and receive priority points.4
The results of our study show that the imaging profile of ICC in the setting of cirrhosis is characterized in most cases by a progressive enhancement after contrast administration. Interestingly, the extent of enhancement is related to tumor size. Hence, in small (≤3 cm) ICCs, the enhancement affects the whole tumor mass, whereas in large (>3 cm) ICCs, the uptake is incomplete, resembling the classic findings of ICCs in the noncirrhotic population.14 A stable enhancement pattern is mostly seen in ICCs smaller than 2 cm. Thereby, 50% of those small ICC nodules exhibit hypersignal in all phases and the other 50% show peripheral enhancement. The first pattern may be observed in a small proportion of small HCCs,8 and if radiologists are not aware of this potential similarity between ICC and HCC, risk is present of false HCC diagnosis.
Progressive contrast enhancement along the different phases of the dynamic study also may be seen in hemangiomas. However, in this entity, the contrast enhancement starts with a characteristic peripheral nodular enhancement, opposite of the rim-like peripheral enhancement of ICC, which precludes its misdiagnosis as ICC.20
Our findings reinforce the need to define the enhancement pattern of any lesion by registering the findings of all dynamic phases after contrast administration. This is key if the aim is to confidently distinguish HCCs from other lesions in cirrhotic liver. Thereby, none of our ICCs displayed contrast washout at delayed phases, a characteristic that is specific to HCC and is the backbone of the noninvasive diagnostic criteria as proposed in the American Association for the Study of Liver Diseases guidelines that have been recently validated.6, 8 The observation of the complete contrast uptake in small nodules within a cirrhotic liver could raise the diagnosis of HCC if there is no awareness of the fact that small ICCS can display this profile that also may be observed in some early HCCs.
Our findings and data from other authors in patients without cirrhosis show that a progressive enhancement pattern is highly suggestive of ICC.11, 12, 21 However, this pattern overlaps with that which may be observed in HCC nodules.8 Accordingly, its identification in the setting of cirrhosis cannot allow a definitive ICC diagnosis by MR imaging. Therefore, if the nodule exhibits a progressive contrast enhancement and absence of washout in delayed phases, independently of its size, biopsy of the lesion should be considered mandatory. Obviously, ICC is highly likely,10–14 but an HCC cannot be discarded. Alpha-fetoprotein will be of no value in this diagnostic effort, because it can be increased in both entities. Similarly, there is no reason to use tumor growth as diagnostic criteria. Both ICC and HCC will exhibit growth as this finding corresponds to malignant entities.
Finally, the characteristic rim-like pattern was only demonstrated in four nodules of 35 mm, 24 mm, 20 mm, and 15 mm. This peripheral rim enhancement may be misread as washout in delayed phases by inexpert radiologists. In our series, the lesions that displayed a peripheral rim-like enhancement in delayed phase showed the same pattern in arterial phase. Moreover, ICC lesions that enhanced homogenously in arterial phase have the same pattern in delayed phase.
Although the pattern of ICC described in our study is well defined in patients without underlying liver disease,4, 10–14 there is limited information about the characteristics of ICC in patients with cirrhosis and, more importantly, when the tumor is detected at an earlier stage. To our knowledge, only Kim and collaborators have intentionally investigated the ICC features on dynamic computed tomography imaging in patients affected with cirrhosis.22 Interestingly, Kim et al. indicated that a significant number of ICC nodules displayed washout, and this could be taken as a potential for a false-positive diagnosis of HCC if applying noninvasive criteria. However, in this study, the mean diameter of the tumors was larger, the imaging technique evaluated was computed tomography, and only arterial and portal phases were described, omitting the valuable information of the delayed phases.
The limitation to our study is the relative small number of reported patients. However, as we previously mentioned, this study analyzed the MR findings of ICC in patients with cirrhosis, and in most cases the nodules were smaller than 3 cm, when radiological diagnosis was a challenge.
In summary, our data provide information that should be of major help in the evaluation of small nodules detected during surveillance in patients with cirrhosis and properly establish an accurate imaging diagnosis. It is important to note that patients without cirrhosis have mostly advanced tumors.
Observation of a washout pattern is specific for HCC, but if this is not observed, a diagnostic biopsy should be considered the sole tool to ensure a conclusive diagnosis.