Intrahepatic mass-forming cholangiocarcinoma and solitary hypovascular liver metastases: is the differential diagnosis using diffusion-weighted MRI possible?

Abstract

Background

Intrahepatic mass-forming cholangiocellular carcinoma (IMC) is the second most common primary liver tumor. The differentiation between IMC and solitary hypovascular liver metastases (SHLM) represents a diagnostic challenge due to many overlapping magnetic resonance imaging (MRI) features.

Purpose

To determine the value of diffusion-weighted imaging (DWI) in addition to conventional MRI for the distinction between intrahepatic mass-forming cholangiocarcinoma and solitary hypovascular liver metastases.

Material and Methods

Fifty-three patients with pathologically proven IMC (n = 31) and SHLM (n = 22) who had undergone MRI and DWI before surgery or percutaneous biopsy were enrolled in this study. The following MRI features were analyzed: the size and shape of the lesion, presence of capsular retraction and segmental biliary dilatation, T2-weighted (T2W) signal intensity, the presence of target sign on DWI and enhancement pattern. Apparent diffusion coefficient (ADC) values were calculated for each lesion (b = 800 s/mm²). Univariate and multivariate logistic regression analyses were used to identify significant differentiating features between IMCs and SHLMs.

Results

Univariate analysis revealed that following parameters favor diagnosis of IMCs over SHLMs: lobulating shape, heterogeneous T2W signal intensity, capsular retraction, segmental biliary dilatation, target sign on DWI and rim-like enhancement on arterial phase followed by progressive enhancement in delayed phases. ADC values measured in the periphery of the lesion were significantly lower in IMCs in comparison to SHLMs. Multivariate analysis revealed that target sign on DWI was the most significant predictor of IMCs.

Conclusion

Qualitative DWI analysis with target sign significantly improves diagnostic accuracy for differentiation among IMC and SHLM lesions.

Keywords

Intrahepatic mass-forming cholangiocarcinoma solitary hypovascular liver lesions magnetic resonance imaging (MRI)diffusion weighted imaging

Introduction

Intrahepatic cholangiocarcinoma is the second most common primary liver tumor that arises from bile duct epithelium, with mass-forming type being the most frequent, accounting for 60% of all these lesions (1,2). Intrahepatic mass-forming cholangiocarcinoma (IMC) usually presents as a large lobulated hypovascular lesion at the time of diagnosis. The most common mimicker of IMC is metastatic adenocarcinoma (3). In the majority of cases the primary origin of the tumor is already known. However, a small subset of patients presents with a metastatic disease for which a primary site is undetectable. The inability to detect the primary tumor could be explained by favorable metastatic ability over local tumor growth or spontaneous tumor regression (4,5). Nevertheless, in 30% of these patients the primary site is eventually detected (6). In addition, patients with known malignant disease could develop IMC independently of their primary disease. Thus, when a patient presents with a hypovascular solitary liver lesion, differentiation among IMC and solitary hypovascular liver metastasis (SHLM) could be a true diagnostic challenge, even on magnetic resonance imaging (MRI). In this regard, several new imaging techniques have been proposed for better detection and characterization of focal liver lesions, such as diffusion-weighted imaging (DWI). DWI provides information on the tissue cellularity based on the detection of the extent of free water molecules diffusion (7). In many organs, apparent diffusion coefficient (ADC) values, as a quantitative measure of diffusion, were observed to be lower in malignant than in benign or normal tissues (8). A variety of malignant hepatic tumors have been studied with DWI (9,10) but only a few studies address the DWI appearance of intrahepatic cholangiocellular carcinoma (11 –13). Of these, only one has comprehensively described DWI features of IMC, both qualitatively and quantitatively (13). In addition, there are limited data regarding differentiation among IMC and SHLM with only one study investigating the role of conventional MRI sequences (14). Therefore, the purpose of this study was to determine MRI features which could improve differential diagnosis among these two types of lesions with the special emphasis on DWI analysis with ADC measurements.

Material and Methods

Study population

This retrospective study was approved by our institutional review board and informed consent was waived. A review of pathology and radiology records between July 2011 and July 2016 was performed to identify eligible patients with following inclusion criteria: (i) patients with pathologically proven IMC (all reports included immunochemistry staining); (ii) patients who underwent MRI according to our routine abdominal protocol prior to surgery; (iii) patients with MRI detectable solitary hypovascular liver metastasis without MRI visible primary tumor; (iv) patients who had no history of prior treatment of the liver. Finally, 31 patient with a diagnosis of IMC (M:F = 18:13; mean age, 52 ± 18 years) and 22 patients with SHLM (M:F = 13:9; mean age 58 ± 16 years) were included. Partial hepatectomy or wedge resection was performed in 35 patients (IMC n = 18, SHLM n = 17) and 18 patients (IMC n = 13, SHLM n = 5) underwent percutaneous biopsy liver biopsy but no further surgery. Thorough review of patient data in follow-up revealed that in ten patients primary tumor was eventually detected with colorectal cancer found in six patients, breast carcinoma in two patients, and lung cancer in two patients.

MRI technique

All patients underwent MRI prior to pathologic diagnosis (mean interval time, 20 days; range, 2–65 days) at 1.5 T (Signa HDxt, GE Healthcare, Waukesha, WI, USA). All images were obtained using an eight-channel phased-array abdominal coil and spine array coil to optimize signal-to-noise ratio (SNR). A rectangular field of view of 320–400 mm was adjusted for each patient’s body size and was held constant for all sequences. Routine breath hold gradient recalled echo (GRE) (in- and out-of-phase) T1-weighted (T1W) sequence, a breath-hold T2-weighted (T2W) single-shot fast spin echo sequence, a breath-hold T2W fat-suppressed (FS), as well as a breath-hold T1W FSGRE sequence were performed. Dynamic imaging was performed after a rapid bolus of a standard dose of 0.1 mmol/kg gadopentetate dimeglumine (Magnevist; Bayer Schering Pharma, Berlin, Germany) at a rate of 1.5 mL/s, immediately followed by a 30 mL saline flush using a power injector. Contrast-enhanced images were acquired with 3D-GRE sequences in the hepatic arterial phase (24 s), portal venous phase (1 min), and interstitial phase (2 min) in the axial plane for all phases, and coronal plane only for interstitial phase. All DWI examinations were obtained with a free-breathing acquisition before intravenous injection of contrast media, using diffusion sensitivity value of 0 and 800 s/mm². The parameters of each pulse sequence are summarized in Table 1.

Table 1.

Parameters of sequences used with 1.5T MRI scanner.

Parameter	T2W single shot FSE	T2W FS	T1W in-phase	T1W out-of phase	T1W 3D-GRE	DW SSSE EPI
TR (ms)	1200	1200	160	160	6.7	5000
TE (ms)	90	90	2	2	4.3	52
Flip angle	90	90	80	80	15	180
BW/pixel (Hz)	62.5	62.5	62.5	62.5	83.33	250
Matrix (phase × frequency)	224 × 288	224 × 288	192 × 256	192 × 256	192 × 320	136 × 136
FOV (mm)	40	40	40	40	40	36
Section thickness (mm)	5	5	5	5	4.4	7
Intersectional gap (%)	20	20	20	20	50	0
Signal acquisition (n)	4	1	1	1	1	3
Fat suppression	None	Fat sat	None	None	Fat sat	None
Respiratory control	BH	BH	BH	BH	BH	RT

3D, three-dimensional; BH, breath-hold; BW, bandwidth; DW, diffusion-weighted; EPI, echo planar imaging; FOV, field of view; FS, fat-suppressed; FSE, fast spin echo; GRE, gradient recalled echo; RT, respiratory triggered; SSSE, single shot spin-echo; TE, echo time; TR, repetition time.

MRI analysis

All MR images were retrospectively reviewed in consensus of two abdominal radiologists (GL and JK with 20 and seven years of experience, respectively) on a picture archiving and communication system workstation (GE ADW4.6; GE Healthcare). Both radiologists were blinded to the pathologic diagnosis, patient’s medical history, and laboratory.

Analysis of morphological features: The following morphological features were analyzed: (i) the size of the lesion; (ii) the location of the lesion; (iii) the shape of the lesion (round or lobulated); (iv) signal intensity (SI) on T2W FS image; (v) the presence of upstream biliary dilatation; (vi) retraction of liver capsule overlying the tumor; and (vii) the presence of lymphadenopathy. The size of each lesion was measured at its greatest diameter in the axial plane on non-contrast T1W images. SI measurements were performed in the lesion and adjacent normal liver parenchyma in T2W FS images. The lesions were categorized as 1 = hypointense (SI lesion < SI liver), 2 = iso-intense (SI lesion = SI liver), 3 = hyperintense (SI lesion > SI liver) compared to the surrounding healthy parenchyma.

Diffusion-weighted imaging characteristics: All lesions were evaluated on high b-value DWI images (b = 800 s/mm²) and classified as: (i) homogeneously hyperintense; (ii) heterogeneously hyperintense; (iii) target appearance (consisting of a central hypointense area and peripheral hyperintense rim; and (iv) isointense. The SI on DWI was assessed in comparison with that of the apparently normal liver parenchyma. Corresponding ADC maps were also qualitatively and quantitatively assessed. Three circular regions of interest (ROI) of 12 ± 3 pixels were manually positioned by two observers on DWI images obtained at b = 0 s/mm²: the first encompassing as much as possible of each lesion, the second ROI was positioned at the tumor periphery, and third ROI was placed in central part of the lesion. ROIs were further transferred using the “copy-and-paste” function from the b = 0 s/mm² DWI image to the ADC map.

Enhancement characteristics on dynamic phase imaging: Analysis of enhancement characteristics on dynamic phase imaging included evaluation of SI of the tumors on each dynamic phase. The contrast enhancement pattern of the tumors depending on a relative enhancement degree in the dynamic phases was classified as: (i) gradual enhancement; (ii) peripheral (rim) enhancement on arterial phase; (iii) mixed type (both rim enhancement and gradual enhancement); and (iv) no or minimal enhancement. Gradual enhancement was defined as an increase in the degree of contrast enhancement within the central parts of the tumor throughout dynamic imaging, irrespective of the early vascular enhancement. The ability to retain contrast was evaluated comparing pre-contrast and arterial phase images with post-contrast images obtained at portal and interstitial phases. No or minimal enhancement was defined as enhancement encompassing less than 50% of the tumor.

Statistical analysis

The univariate statistical differences between each MRI parameter were analyzed using a chi-square test or Fisher’s exact test for nominal variables. Significant variables (P < 0.05) at univariate analysis were used as input variables for multivariate logistic regression analysis to determine the most relevant findings for distinction among IMCs and SHLMs. For calculation of sensitivity and specificity descriptive statistics and crosstabs option in statistical program were used with histopathological diagnosis as gold standard. For that purpose, all IMC lesions were divided into two groups, the first group included IMCs, which displayed all significant imaging features from multivariate analysis and the second group included all the other IMCs. Analysis of variance (ANOVA) was used to compare continuous variables between IMCs and SHLMs. Statistical significance level was set at 0.05. All analyses were performed with software SPSS (version 15.0 for Windows; SPSS, Chicago, IL, USA).

Results

Morphological features

The mean size of IMC lesions was 7.4 ± 3.2 cm, which was significantly higher (P = 0.031) in comparison to SHLMs (mean 3.8 ± 1.6 cm). A heterogeneous T2W SI, lobulated contour, capsular retraction, segmental biliary dilatation upstream the tumor, and presence of central T2W hypointensity were significant predictors of IMCs (all, P < 0.05) (Fig. 1). The incidence of lymphadenopathy did not differ significantly among IMCs and SHLMs (P = 0.598) While enlarged lymph nodes in IMCs were mostly present in porta hepatis and hepatoduodenal space, in the SHLM group para-aortal lymphadenopathy was the most common pattern. Detailed description of morphological MRI features of both IMCs and SHLMs is provided in Table 2.

Fig. 1.

MRI features of intrahepatic mass-forming cholangiocarcinoma. (a) Axial T2W FS image shows lobulated heterogeneously slightly hyperintense lesion occupying segments IVb and V in a 53-year-old man with pathologically confirmed IMC. Note also irregular area of central hypointensity corresponding to dense fibrosis (asterisk). (b) Coronal T2W FS image shows slightly hyperintense lesion in segment IV with biliary dilatation upstream the tumor and overlying capsular retraction (arrow).

Table 2.

Morphological MRI features of IMCs and SHLM.

Variables	IMC (n = 31)	SHLM (n = 22)	P value
Size (cm)
>5	28 (90.3%)	2 (9.1%)	<0.05*
T2W FS SI
Homogeneous	0 (0%)	16 (72.4%)	<0.001*
Heterogeneous	31 (100%)	6 (27.3%)	<0.05
Central hypointense area	8 (25.8%)	0 (0%)	<0.001*
Capsular retraction	24 (77.4%)	0 (0%)	<0.001*
Segmental bile duct dilatation	7 (22.6%)	0 (0%)	<0.001*
Lymphadenopathy	20(64.5%)	10 (45.4%)	<0.05
Shape
Lobulated	31 (100%)	8 (36.4%)	<0.05*
Round	0 (0%)	14 (63.6%)	<0.05*
Target sign on DWI	29 (93.5%)	1 (4.5%)	<0.001*

Data are numbers of patients with percentages in parentheses.

Data are statistically significant.

DWI, diffusion-weighted imaging; FS, fat-suppressed; IMC, intrahepatic mass-forming cholangiocarcinoma; SHLM, solitary hypovascular liver metastasis; SI, signal intensity.

DWI characteristics

Target sign on DWI was present in 29 ICCs (93.5%) and in only one patient with SHLM (P = 0.0006) (Figs 2 and 3). No significant difference was found among ADC values calculated for ROI encompassing whole IMC lesion versus SHLM (P = 0.245). However, the ADC values measured on the edge of IMCs were slightly but significantly smaller than ADC values of the same part of SHLM group (P = 0.037). Detailed description of ADC values of both groups is provided in Table 3.

Fig. 2.

DWI features of intrahepatic mass-forming cholangiocarcinoma. (a) High b-value DWI shows target appearance of intrahepatic mass-forming cholangiocarcinoma consisting of central darker area and peripheral hyperintense area with corresponding ADC map in a 57-year-old man. (b) Target sign on axial DWI and ADC map in a 66-year-old woman with IMC.

Fig. 3.

A 74-year-old man with solitary hypovascular liver metastasis. (a) Axial T2W FS image shows homogeneously hyperintense lesion in liver segment VI with bulging of the overlying liver capsule (arrow). (b) On the axial diffusion-weighted imaging at b = 800 s/mm², the tumor shows homogeneously high signal intensity. (c) Corresponding ADC map shows uniformly restricted diffusion in whole lesion.

Table 3.

ADC values of IMC and SHLM.

	IMC (n = 31)	SHLM (n = 22)	P value
ADC (ROI encompassing whole lesion)	1.001 ± 0.112	1.046 ± 0.036	>0.05
ADC measured at tumor periphery	0.813 ± 0.221	1.021 ± 0.187	<0.05*
ADC measured in central part of the tumor	1.325 ± 0.128	1.079 ± 0.238	<0.05*

Results are expressed as 10^–3 mm²/s.

Data are statistically significant.

ADC, apparent diffusion coefficient; IMC, intrahepatic mass-forming cholangiocarcinoma; ROI, region of interest; SHLM, solitary hypovascular liver metastasis.

Enhancement characteristics

On dynamic imaging the most common pattern of enhancement in IMCs was rim enhancement on arterial phase with progressive filling on portal venous and interstitial phase, seen in 23 (74.2%) of patients which was slightly but significantly different from SHLM group (P = 0.021) (Fig. 4a). In three IMC patients, no or minimal enhancement was noted. Complete enhancement was observed in only six (19.3%) IMC patients (Fig. 4b), while in the rest a central hypointense area was seen on delayed images (Fig. 4c). The most common enhancement pattern found in SHLMs was rim enhancement while none of the IMC patients displayed only rim enhancement (P = 0.0007) (Fig. 5). Table 4 describes the enhancement characteristics of both groups in greater detail.

Fig. 4.

Three different cases of intrahepatic mass-forming cholangiocarcinoma. Images were obtained with T1W sequences before contrast administration, in arterial phase, portal phase and interstitial phase (a–c). (a) Typical pattern of IMC with peripheral rim-like enhancement in arterial phase (arrow), followed by progressive filling on the portal venous and interstitial phase. (b) Early inhomogeneous enhancement on arterial phase, followed by prolonged almost complete enhancement on the late phases. (c) Pattern with early peripheral rim enhancement on arterial phase with non-filling of the central area on the delayed images.

Fig. 5.

Three different cases of solitary hypovascular liver metastases (a–c). Images were obtained with T1W sequences before contrast administration, in arterial phase, portal phase and interstitial phase. (a) Typical pattern of SHLM with ring enhancement in arterial phase. Delayed phase images show fading of the outer portion with subtle central retention of contrast. Note also multiple simple liver cysts surrounding metastasis located in segment VIII. (b) Ring enhancement on arterial phase and fading of the outer portion to near isointensity on portal and interstitial phase in SHLM lesion located in segment VIII. (c) Pattern with early peripheral enhancement on arterial phase with progressive filling on portal and interstitial phase.

Table 4.

Enhancement characteristics of IMC and SHLM.

Enhancement pattern	IMC (n = 31)	SHLM (n = 22)	P value
Rim enhancement on AP	0 (0%)	13 (59.1%)	<0.001*
Gradual enhancement	5 (16.1%)	4 (18.2%)	>0.05
Mixed enhancement	23 (74.2%)	1 (4.5%)	<0.001*
No or minimal enhancement	3 (13.6%)	4 (18.2%)	>0.05

Data are statistically significant.

Data are numbers of patients with percentages in parentheses.

AP, arterial phase; IMC, intrahepatic mass-forming cholangiocarcinoma; SHLM, solitary hypovascular liver metastasis.

Multivariate analysis

Multivariate binary logistic regression analysis showed that target sign on DWI was the most significant predictor for differentiation of IMC and SHLM lesions (P = 0.0006), followed by capsular retraction, segmental biliary dilatation, and heterogeneous T2W FSSI . If all three imaging features were used, IMC could be diagnosed with 92% sensitivity and 81% specificity (Table 5).

Table 5.

Results of multivariate logistic regression analysis of variables for differentiation among IMC and SHLM.

	Logistic regression coefficient	Standard error	P value	Odds ratio	95% CI of odds ratio
Target sign	4.43	1.36	0.0006	91.41	8.65–1001.87
Capsular retraction	3.78	1.23	0.0009	62.65	3.89–411.90
Segmental biliary dilatation	2.65	1.19	0.005	22.98	2.28–301.67
Heterogeneous T2W SI	2.14	1.05	0.014	16.76	1.87–142.77

CI, confidence interval; SI, signal intensity.

Discussion

The results from the present study revealed that addition of DWI with ADC measurements to conventional MRI protocol facilitates distinction between the two most common solitary hypovascular liver lesions, IMC and SHLM. The assessment made using only conventional MRI sequences does not provide accurate differential diagnosis since there are many overlapping features (14). Moreover, the differentiation among IMCs and SHLMs may be even histologically very difficult, since some metastatic liver tumors, especially from pancreatic and gallbladder cancers, have similar cytokeratin expression as IMC (15,16). Thus, in the case where MRI findings suggest diagnosis of SHLM, the thorough search for primary tumor should be undertaken. Although difficult, the distinction between IMC and SHLM is of clinical importance from a prognostic and therapeutic standpoint, as the only potentially curative treatment for IMC is complete surgical resection with lymphadenectomy, while SHLM diagnosis requires more individual approach with surgical resection in some cases and chemotherapy as concomitant or the only treatment in other patients (17,18).

We found the target sign, defined as a central hypointensity on high-b-value DWI with a peripheral hyperintense rim (12), in 29 (93.5%) IMC lesions while only one (4.5%) patient with SHLM showed this feature. Furthermore, all IMCs displayed the target sign on ADC map regardless their appearance on DWI. Similarly, Park et al. (12) found the target sign to be a significant predictor of IMC as it was present in 75% of IMC patients while none of HCC lesions showed this feature. Target appearance seen on DWI was considered to be correlated to pathological features of IMCs (19,20). Namely, the majority of IMCs contain loose central fibrosis with accompanying edema, which is responsible for low signal intensity on high-b-value DWI (21) while the periphery of the tumor is composed of densely packed viable cells which cause diffusion restriction and dark ring on ADC map (22). In addition, IMCs commonly show target appearance on the hepatobiliary phase, after administration of hepatospecific contrast agents, like Gd-EOB-DTPA or Gd-BOPTA, as well as on DWI. This multilayered pattern on hepatobiliary phase is also explained by pathological composition of IMCs with greater cellularity and vascularity in the periphery of the tumor, which release contrast in the late phases, and rich fibrous stroma in central parts which retain contrast in extracellular space. The results from the present study revealed that among all other characteristics, the target sign was the most significant predictor of IMCs. When it was combined with other significant features from univariate analysis, a sensitivity of 92% and a specificity of 81% was reached. Therefore, we could assume that qualitative analysis of DWI images should always be a part of the MRI protocol when evaluating hypovascular liver lesions.

In our study, we also performed a quantitative analysis of IMCs and SHLMs in terms of ADC measurements. To our knowledge there is only one study in previous literature describing ADC values of intrahepatic mass-forming cholangiocarcinoma (13). Thus Fattach et al. (13) in their study including 11 IMCs, found a mean ADC of 1.042 × 10^–3 mm²/s at 1.5 T using three b values (0, 400, 800 s/mm²), which is similar to the results of the present study. Comparing ADC values calculated for ROI encompassing whole lesion, we found no significant difference among IMCs and SHLMs. Nevertheless, when comparing ADC values from central tumor parts and periphery of the lesion, a significant difference was found with the lowest values at the edge of the IMCs (0.813 versus 1.325 × 10^–3 mm²/s), while it was not observed in SHLMs. These results correlate with the high incidence of target sign in IMCs and probably correspond to the high cell density at the periphery and less restricted diffusion in central parts while SHLM lesions seem to be more uniformly hypercellular (23,24).

Several other imaging features could be useful for distinction among IMCs and SHLMs, such as shape of the lesion, capsular retraction, segmental bile duct dilatation, and portal lymphadenopathy. All IMCs in our study presented as large lobulated masses, in contrast to SHLMs which were mostly (14/22, 63.6%) round lesions with smooth margins. However, this sign is not sufficient for distinction among these two types of tumors since it is widely known that colorectal liver metastases may also present as lobulated lesions with wavy contours (25). In our series, 77.4% (24/31) of IMCs showed overlying capsule retraction sign in accordance with previous reports, for instance 56% in the study by Ansari et al. (14). Nevertheless, since this finding could be also found in other liver tumors, it is not specific for IMCs but only suggestive of a malignant tumor with a relatively prominent desmoplastic reaction (26). Similarly, even though segmental biliary dilatation directly related to the tumor is a frequent finding in IMCs, as found in seven patients (22.6%) of our study population, it could also be present in metastatic lesions due to the compression of the tumor on biliary tree (27). The absence of this sign in SHLMs in the present study could be attributed to the relatively small number of these patients and to the small size of metastatic tumors (mean 3 cm in diameter). Porta hepatis lymphadenopathy was previously described as one of the expanded criteria for IMCs diagnosis (14). Consistent with these findings, our results have shown that enlarged lymph nodes in porta hepatis were seen only in IMC patients, while patients with metastatic tumors had paraaortic lymphadenopathy. However according to Ishida et al. (28) portal lymphadenopathy could also occur in metastatic lesions due to lymphatic drainage and thus could not be consider sensitive only for ICC.

As previously described (29,30), the most common dynamic enhancement pattern of IMCs is minimal to moderate thin peripheral rim enhancement on early images with progressive centripetal or discontinuous filling on delayed images, found in 74.2% (23/31) of our patients. This behavior might be attributed to rich fibrous stroma found in mass-forming intrahepatic cholangiocarcinoma and slow diffusion of contrast material into the interstitial spaces of the tumor (2,30). In our study, all lesions with peripheral rim enhancement also showed the target sign on DWI. Peripheral hyperintensity on DWI configuring the “target sign” is likely correlated with the hypervascular peripheral rim in the arterial phase, as both findings correspond to the greater density of viable and vascularized tumor cells in the periphery of the tumor. On the contrary, the most common pattern in SHLM group was rim enhancement in arterial phase with no or minimal central enhancement on delayed images, consistent with previous reports (4). As metastatic tumors develop from cells which do not originate from the liver, they parasitize the surrounding blood vessels, creating the rim appearance of blood supplying the most vascularized outer parts of the tumor (31). It is interesting to note that four patients (18.2%) in our study with SHLM demonstrated gradual enhancement similar to IMCs, which was also observed in 17% of metastatic lesions in the study by Danet et al. (31). Therefore, on the basis of our results, it could be concluded that even this type of enhancement is considered characteristic for IMC, it is not sufficiently specific for accurate differentiation among these two types of hypovascular liver lesions.

Our study has several limitations. First, it was limited by its retrospective design which may have introduced inherent selection bias. Second, none of the patients in our study population had small IMC lesion (<3 cm), which could display atypical enhancement pattern due to smaller amount of intratumoral fibrosis. Therefore, the incidence and significance of target sign for these lesions could not be assessed. In addition, all SHLMs evaluated in this study were small lesions (around 3 cm), homogeneously T2W hyperintense lesions without areas of central necrosis which could create an appearance similar to the target sign. Third, in our study we did not use a hepatospecific contrast agent which could provide additional information for differentiation among these two types of hypovasular liver lesions.

In conclusion, even though IMCs and SHLMs show many similar features, the target sign on DWI and ADC map was shown to be the most significant predictor among these two types of hepatic tumors. Furthermore, ADC measurements at the periphery and central parts of the tumor could facilitate differential diagnosis since they are significantly different in IMCs in comparison to SHLM lesions.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Choi

Lee

Han

. Imaging of intrahepatic and hilar cholangiocarcinoma. Abdom Imaging 2004; 29: 548–557.

Lim

. Cholangiocarcinoma: morphologic classification according to growth pattern and imaging findings. Am J Roentgenol 2003; 181: 819–827.

Lee

Lim

Jang

et al.

Radiologic spectrum ofcholangiocarcinoma: emphasis on unusual manifestations and differential diagnosis. Radiographics 2001; 21: 97–116.

Braga

Semelka

Danet

et al.

Liver metastases from unknown primary site: demonstration on MR images. Magn Reson Imaging 2003; 21: 871–877.

Abbruzzese

Lenzi

Raber

et al.

The biology of unknown primary tumors. Semin Oncol 1993; 20: 238–243.

Pavlidis

Pentheroudakis

. Cancer of unknown primary site. Lancet 2012; 379: 1428–1435.

Taouli

Koh

. Diffusion-weighted MR imaging of the liver. Radiology 2010; 254: 47–66.

Koh

Collins

. Diffusion-weighted MRI in the body: applications and challenges in oncology. Am J Roentgenol 2007; 188: 1622–1635.

Kanematsu

Goshima

Watanabe

et al.

Detection and characterization of focal hepatic lesions with diffusion-weighted MR imaging: a pictorial review. Abdom Imaging 2013; 38: 297–308.

10.

Taouli

. Diffusion-weighted MR imaging for liver lesion characterization: a critical look. Radiology 2012; 262: 378–380.

11.

Park

Kim

Park

et al.

Scirrhous hepatocellular carcinoma on gadoxetic acid-enhanced magnetic resonance imaging and diffusion-weighted imaging: emphasis on the differentiation of intrahepatic cholangiocarcinoma. J Comput Assist Tomogr 2013; 37: 872–881.

12.

Park

Kim

Park

et al.

Small intrahepatic mass-forming cholangiocarcinoma: target sign on diffusion-weighted imaging for differentiation from hepatocellular carcinoma. Abdom Imaging 2013; 38: 793–801.

13.

Fattach

Dohan

Guerrache

et al.

Intrahepatic and hilar mass-forming cholangiocarcinoma: Qualitative and quantitative evaluation with diffusion-weighted MR imaging. Eur J Radiol 2015; 84: 1444–1451.

14.

Al Ansari

Kim

Srirattanapong

et al.

Mass-forming cholangiocarcinoma and adenocarcinoma of unknown primary: can they be distinguished on liver MRI?

Abdom Imaging 2014; 39: 1228–1240.

15.

Shimonishi

Miyazaki

Nakanuma

. Cytokeratin profile relates to histological subtypes and intrahepatic location of intrahepatic cholangiocarcinoma and primary sites of metastatic adenocarcinoma of liver. Histopathology 2000; 37: 55–63.

16.

Somorácz

Tátrai

Horváth

et al.

Agrin immunohistochemistry facilitates the determination of primary versus metastatic origin of liver carcinomas. Hum Pathol 2010; 41: 1310–1319.

17.

Valverde

Bonhomme

Farges

et al.

Resection of intrahepatic cholangiocarcinoma: a Western experience. J Hepatobiliary Pancreat Surg 1999; 6: 122–127.

18.

Bridgewater

Galle

Khan

et al.

Guidelines for the diagnosis and management of intrahepatic cholangiocarcinoma. J Hepatol 2014; 60: 1268–1289.

19.

Kim

Lee

Shin

et al.

Differentiation of intrahepatic mass-forming cholangiocarcinoma from hepatocellular carcinoma on gadoxetic acid-enhanced liver MR imaging. Eur Radiol 2016; 26: 1808–1817.

20.

Jeong

Kim

Chung

et al.

Gadoxetate disodium-enhanced MRI of mass-forming intrahepatic cholangiocarcinomas: imaging-histologic correlation. Am J Roentgenol 2013; 201: W603–611.

21.

Maetani

Itoh

Watanabe

et al.

MR imaging of intrahepatic cholangiocarcinoma with pathologic correlation. Am J Roentgenol 2001; 176: 1499–1507.

22.

Okamura

Yoshida

Shibuya

et al.

Cellular and stromal characteristics in the scirrhous hepatocellular carcinoma: comparison with hepatocellular carcinomas and intrahepatic cholangiocarcinomas. Pathol Int 2005; 55: 724–731.

23.

Kim

Lee

et al.

Diagnostic accuracy and sensitivity of diffusion-weighted and of gadoxetic acid-enhanced 3-T MR imaging alone or in combination in the detection of small liver metastasis (≤ 1.5 cm in diameter). Invest Radiol 2012; 47: 159–166.

24.

Kim

Han

Kim

. Comparison of diffuse hepatocellular carcinoma and intrahepatic cholangiocarcinoma using sequentially acquired gadolinium-enhanced and Resovist-enhanced MRI. Eur J Radiol 2009; 70: 94–100.

25.

Outwater

Tomaszewski

Daly

et al.

Hepatic colorectal metastases: correlation of MR imaging and pathologic appearance. Radiology 1991; 180: 327–332.

26.

Soyer

Bluemke

Vissuzaine

et al.

CT of hepatic tumors: prevalence and specificity of retraction of the adjacent liver capsule. Am J Roentgenol 1994; 162: 1119–1122.

27.

Jhaveri

Halankar

Aguirre

et al.

Intrahepatic bile duct dilatation due to liver metastases from colorectal carcinoma. Am J Roentgenol 2009; 193: 752–756.

28.

Ishida

Ishibashi

Ohsawa

et al.

Significance of hepatic lymph node metastasis in patients with unresectable synchronous liver metastasis of colorectal cancer. Int Surg 2011; 96: 291–299.

29.

Chung

Kim

Park

et al.

Varying appearances of cholangiocarcinoma: radiologic-pathologic correlation. Radiographics 2009; 29: 683–700.

30.

Kim

Lee

et al.

Intrahepatic mass-forming cholangiocarcinomas: enhancement patterns at multiphasic CT, with special emphasis on arterial enhancement pattern–correlation with clinicopathologic findings. Radiology 2011; 260: 148–157.

31.

Danet

Semelka

Leonardou

et al.

Spectrum of MRI appearances of untreated metastases of the liver. Am J Roentgenol 2003; 181: 809–817.