Combined Hepatocellular and Cholangiocarcinoma (Biphenotypic) Tumors: Imaging Features and Diagnostic Accuracy of Contrast- Enhanced CT and MRI

Transcription

1 Gastrointestinal Imaging Original Research Fowler et al. Imaging Biphenotypic Tumors Gastrointestinal Imaging Original Research Kathryn J. Fowler 1 Arman Sheybani 2 Rex A. Parker, III 3 Sean Doherty 1 Elizabeth M. Brunt 4 William C. Chapman 5 Christine O. Menias 1 Fowler KJ, Sheybani A, Parker RA III, et al. Keywords: biphenotypic, cholangiocarcinoma, CT, hepatocellular carcinoma, MRI DOI: /AJR Received June 22, 2012; accepted after revision October 1, K. J. Fowler has received research support from Bracco Diagnostics for a separate project and is on the speakers bureau for Lantheus Medical Imaging. 1 Department of Radiology, Division of Abdominal Imaging, Washington University School of Medicine in St. Louis, Mail Stop 8131, 510 S. Kingshighway Blvd, St. Louis, MO Address correspondence to K. J. Fowler (fowlerk@mir.wustl.edu). 2 Department of Medicine, Icahn School of Medicine at Mount Sinai, New York, NY. 3 Kaiser Permanente Medical Center, Los Angeles, CA. 4 Department of Pathology, Washington University, St. Louis, MO. 5 Department of Surgery, Washington University, St. Louis, MO. AJR 2013; 201: X/13/ American Roentgen Ray Society Combined Hepatocellular and Cholangiocarcinoma (Biphenotypic) Tumors: Imaging Features and Diagnostic Accuracy of Contrast- Enhanced CT and MRI OBJECTIVE. The purpose of this study was to evaluate the diagnostic accuracy of preoperative imaging for diagnosis of combined and to evaluate the clinical and imaging features and demographics of patients presenting to our institution with such tumors. MATERIALS AND METHODS. From January 2001 to January 2011, 29 patients presented with pathologically proven combined to our institution. A retrospective review of the imaging studies, clinical data, and demographic information in these patients was conducted. Two radiologists with 6 and 18 years of experience reviewed the imaging studies of patients with tumors and matched control cases of hepatocellular carcinoma (HCC) (n = 15) and cholangiocarcinoma (n = 18). The reviewers were blinded to the pathologic diagnosis. Imaging features on contrast-enhanced MRI and CT with the suggested final diagnosis were recorded. RESULTS. The demographics of our patient population were similar to other reported U.S. populations, with cirrhosis and hepatitis present in a minority of patients. The imaging features of combined overlapped with those of HCC and cholangiocarcinoma. The correct diagnosis of tumors was made in a minority of cases by either radiologist, with misdiagnosis more often leading to suggestion of cholangiocarcinoma than HCC. Sensitivities and specificities for diagnosis of combined ranged from 33% to 34% and 81% to 100%, respectively. CONCLUSION. Preoperative diagnosis of tumors on the basis of imaging features is accurate in the minority of cases. Tumor markers and risk factors may help improve accuracy; however, in the absence of classic imaging features and supportive information for HCC or cholangiocarcinoma, biopsy should be considered for confirmation of diagnosis. C holangiocarcinoma and hepatocellular carcinoma (HCC) are the most common primary liver malignancies and are derived from cholangiocytes and hepatocytes, respectively. Combined HCC and cholangiocarcinoma tumors, also referred to as biphenotypic tumors, comprise a minority of primary hepatic malignancies, accounting for % [1]. Initially described by Allen and Lisa [2] in 1949, combined hepatocellular cholangiocarcinoma tumors have been categorized into three distinct types on the basis of the relative separation of the different cholangiocarcinoma and HCC components. Type C, consisting of intermixed components and transitional cell types appears to represent a true combination tumor whereas other types may exist along the spectrum of collision tumors [3, 4]. It is theorized that evolution of tumors from common hepatic progenitor cells may account for the biphenotypic features of these tumors [5 12]. The expression of multiple progenitor cell markers by scirrhous HCC, a rare fibrous subtype of HCC that shows intermediate traits between HCC and cholangiocarcinoma also lends support to this theory of evolution from a common progenitor cell [13]. With few exceptions, studies suggest risk factors and patient demographics are similar for tumors, HCC, and cholangiocarcinoma and include advanced age, male predominance, hepatitis B viral infection, and 332 AJR:201, August 2013

2 Imaging Biphenotypic Tumors TABLE 1: Demographic Information for Patients With Combined Hepatocellular Cholangiocarcinoma Tumors and Control Subjects Parameter Combined Hepatocellular Cholangiocarcinoma Tumors Cholangiocarcinoma HCC Mean age ± SD (y) 61.2 ± ± ± 14.5 Age range (y) Sex (M/F) 13/16 9/9 7/8 Mean AFP (ng/ml) 80.9 (0) 3.9 (3.8) 4640 (18.2) Mean CA 19-9 (ng/ml) 2,770 (39.8) 64,150 (214) 18.6 (13.0) Hepatitis B a Hepatitis C b Cirrhosis c Alcoholic cirrhosis Hereditary hemochromatosis Primary sclerosing cholangitis Note Data in parentheses are medians. HCC = hepatocellular carcinoma, AFP = α-fetoprotein, CA = carbohydrate antigen. a Positive serology for hepatitis B core antigen. b Positive serology for hepatitis C. c Risk factors in patients with cirrhosis (alcoholic cirrhosis, hereditary hemochromatosis, primary sclerosing cholangitis, or hepatitis C). chronic liver disease [3, 4, 14 19]. However, the relative rarity of combined hepatocellular cholangiocarcinoma tumors, geographic heterogeneity in incidence, and small patient populations reported in most series confounds clear understanding of the demographic features and clinical behavior. Additionally, the mixed pathologic nature and varying subtypes of tumors also lends heterogeneity to study populations, making comparisons across case series somewhat difficult. Given the overlap in patient demographics and the unique nature of combined hepatocellular cholangiocarcinoma tumors, preoperative diagnosis is crucial for appropriate management. This is especially true in the United States, where HCC is diagnosed on the basis of imaging appearance alone, which then in turn determines transplantation exception status and management decisions. The prognosis of combined hepatocellular cholangiocarcinoma tumors compared with cholangiocarcinoma and HCC remains controversial, with most studies showing slightly worse prognosis and a few showing intermediate survival [3, 4, 15, 16, 18, 20 27]. A potential confounding issue in understanding survival and treatment effect in patients with tumors relates to the retrospective nature of most studies, in which it is unclear if the biphenotypic pathology of the malignancy was known preoperatively and whether a different definitive treatment may have provided survival or time to progression benefit. Given the rarity of this tumor type, appropriate management remains controversial. Surgical resection likely yields the greatest survival benefit in limited-stage patients. To our knowledge, few studies have evaluated the imaging characteristics of combined and no studies have evaluated the ability of preoperative imaging to determine diagnosis. The appearance of HCC and cholangiocarcinoma tumors is well known on contrast-enhanced MRI and CT. The histologic composition and relative ratio of cholangiocarcinoma and HCC components within combined hepatocellular cholangiocarcinoma tumors appear to dictate the imaging appearance. Tumors may show features typical of HCC, such as arterial enhancement, washout, and pseudocapsule, whereas other regions within the tumor show progressive or delayed enhancement, necrosis, and possible ductal dilation more akin to cholangiocarcinoma [3, 28 31]. Some suggest that the combination of imaging features and tumor markers may be helpful in preoperative diagnosis of combined hepatocellular cholangiocarcinoma tumors [3]. The main tumor markers of interest are carbohydrate antigen (CA) 19-9 and α fetoprotein (AFP), which are useful adjuncts to imaging in patients with cholangiocarcinoma and HCC, respectively. Although AFP alone may be a relatively insensitive tumor marker, simultaneous elevation of both CA 19-9 and AFP has been suggested as highly concerning for combined hepatocellular cholangiocarcinoma tumors [27]. Other reports suggest that discordance between tumor marker elevation and imaging morphology may be suggestive [3]. For instance, a tumor that resembles a cholangiocarcinoma on imaging but seen in combination with elevated AFP may suggest a biphenotypic tumor. The purpose of this study was to retrospectively evaluate the demographics, clinical presentation, and imaging features in patients in whom the pathologic diagnosis of tumors had been made by biopsy or at explantation. This is of particular interest given the relative lack of representation of Western populations in the literature and unclear role of preoperative imaging in establishing a diagnosis of combined hepatocellular cholangiocarcinoma tumors. Our goal was to elucidate the imaging features of combined on dynamic contrast-enhanced MRI and CT in the hope of defining features that may enable preoperative diagnosis and better guide clinical management decisions. Materials and Methods Patient Population and Clinical Information After institutional review board approval, all cases of tumors primary hepatic malignancy were identified by retrospective search of the radiology and pathology databases from January 2001 to January All patients with available crosssectional imaging were included in the study; 10 AJR:201, August

3 Fowler et al. TABLE 2: Imaging Features of Lesion Types on MRI and CT by Reviewers Parameter Arterial Enhancement patients without available imaging were excluded. The patient population consisted of 29 patients with tumors: 13 men and 16 women. Additional control cases were identified, consisting of cholangiocarcinoma (n = 18) and HCC (n = 15), which were matched for age, presence of cirrhosis, and relative size of the mass in question. Clinical information from these 62 patients is presented in Table 1. Tumor markers reported were drawn before initiating any therapy and within 1 week of the imaging study. Cirrhosis had been confirmed by biopsy. Survival and treatment data were obtained by chart review. Because not all patients underwent complete resection, tumor size is presented as the largest dimension on cross-sectional imaging. Imaging Protocols and Analysis All patients with tumors had undergone contrast-enhanced cross-sectional imaging, either MRI (n = 21) or CT (n = 8). Control cases included both imaging modalities, MRI (n = 20) and CT (n = 13). The Washout Delayed Enhancement Capsular Retraction Biliary Duct Dilation Pseudocapsule Fat Combined hepatocellular cholangiocarcinoma tumors (n = 29) Reviewer 1 overall (%) 100 (27/27) 41 (11/27) 48 (13/27) 42 (12/29) 35 (10/29) 26 (7/27) 0 (0/29) CT (8) MRI (21) Reviewer 2 overall (%) 100 (27/27) 33 (9/27) 74 (20/27) 45 (13/29) 17 (5/29) 22 (6/27) 0 (0/29) CT (8) MRI (21) Hepatocellular carcinoma (n = 15) Reviewer 1 overall (%) 92 (12/13) 100 (13/13) 8 (1/13) 15 (2/13) 7 (1/15) 46 (6/13) 20 (3/15) CT (8) MRI (7) Reviewer 2 overall (%) 100 (13/13) 100 (13/13) 15 (2/13) 33 (5/15) 7 (1/15) 46 (6/13) 40 (6/15) CT (8) MRI (7) Cholangiocarcinoma (n = 18) Reviewer 1 overall (%) 76 (13/17) 18 (3/17) 100 (17/17) 72 (13/18) 83 (15/18) 6 (1/17) 0 (0/18) CT (5) MRI (13) Reviewer 2 overall (%) 82 (14/17) 6 (1/17) 94 (16/17) 56 (10/18) 67 (12/18) 0 (0/17) 0 (0/18) CT (5) MRI (13) Note Numbers in parentheses were used to generate percentages. Numbers reflect number of patients (percentages) with denominators reflecting number of multiphase studies for enhancement features and total number of patients for other features. Number of cases differs from total because only cases with dynamic multiphase contrast-enhanced images were included for features relating to enhancement patterns. MRI protocols had been performed with standard sequences, including T2-weighted fat-suppressed and non-fat-suppressed axial, T1-weighted in- and opposed-phase, and unenhanced and dynamic contrastenhanced 3D gradient-recalled echo fat-suppressed imaging. Seven patients underwent hepatobiliary phase MRI with either gadobenate dimeglumine (MultiHance, Bracco Diagnostics) or gadoxetic acid (Eovist, Bayer HealthCare). All CT examinations were performed using a liver protocol with multiphase dynamic studies consisting of arterial, portal venous, and delayed phase imaging, except for two of the tumors cases, two HCC cases, and a single cholangiocarcinoma case, which included the portal venous phase only. Features requiring multiphase assessment, such as arterial enhancement, washout, and pseudocapsule, were not assessed for these studies. All available imaging was reviewed again as part of the final assessment by the radiologists. The imaging studies for the patients with combined and control cases were reviewed on a PACS workstation by two fellowship-trained abdominal radiologists (reviewer 1 with 6 years of experience and reviewer 2 with 18 years of experience). The study radiologists were informed that the patient had either tumors, HCC, or cholangiocarcinoma but were blinded to the final diagnosis. The presence or absence of arterial enhancement, washout, pseudocapsule, delayed (progressive) enhancement, diffusion restriction, biliary dilatation, tumoral fat (defined by loss of signal on chemical-shift imaging), capsular retraction or fibrosis, T1/T2 signal intensity, and morphology of arterial enhancement (homogeneous, heterogeneous, or peripheral) were recorded. Targetlike enhancement with initial peripheral arterial enhancement and central hypoenhancement followed by progressive central fill-in and peripheral washout was assessed. Additionally, background liver cirrhosis, steatosis, iron deposition, and vascular invasion were noted. The presence of steatosis or iron deposition was suggested on the basis of chemical-shift imaging on MRI or abnormal attenuation of the liver on CT compared 334 AJR:201, August 2013

4 Imaging Biphenotypic Tumors TABLE 3: Enhancement Characteristics of Different Tumor Types Arterial Enhancement Characterization Parameter with the spleen. The radiologists were also asked to make a most likely diagnosis on the basis of the imaging features of the lesion. None Peripheral Homogeneous Heterogeneous Tissue Diagnosis Histopathologic diagnosis of combined hepatocellular cholangiocarcinoma tumors was by surgical resection in 12 patients, transplantation in two patients, and core biopsy in 15 patients. In patients with cholangiocarcinoma, biopsy or surgical resection was the final source of tissue for histopathologic evaluation and final diagnosis. All patients with HCC underwent resection and tissue diagnosis of the final surgical specimen. Pathology reports were reviewed for final histologic diagnosis (combined, HCC, or cholangiocarcinoma), underlying cirrhosis or fibrosis, presence of lymphovascular invasion, positive lymph nodes, and satellite lesions. The diagnosis of biphenotypic primary liver carcinoma depended on a combination of H and E stain findings and final proof of both hepatocellular and biliary differentiation by immunohistochemical markers. The former include polyclonal carcinoembryonic antigen, a marker of canalicular formation, and the latter, keratin 7 or keratin 19. Subclassification, when possible, was done according to the 2010 World Health Organization Tumours of the Digestive System classification [2]. Reverse Target Appearance Combined hepatocellular cholangiocarcinoma tumors (n = 27) Reviewer 1 overall (%) 0 (0) 16 (59) 5 (19) 6 (22) 10 (37) CT (6) MRI (21) Reviewer 2 overall (%) 0 (0) 14 (52) 2 (7) 11 (41) 11 (41) CT (6) MRI (21) Hepatocellular carcinoma (n = 13) Reviewer 1 overall (%) 2 (15) 1 (8) 4 (31) 6 (46) 0 (0) CT (6) MRI (7) Reviewer 2 overall (%) 0 (0) 2 (15) 1 (8) 10 (77) 0 (0) CT (6) MRI (7) Cholangiocarcinoma (n = 17) Reviewer 1 overall (%) 4 (23) 10 (59) 0 (0) 3 (18) 9 (69) CT (5) MRI (13) Reviewer 2 overall (%) 2 (12) 11 (65) 0 (0) 4 (23) 10 (77) CT (5) MRI (13) Note Reviewers characterized the pattern of arterial enhancement on multiphase imaging examinations for each tumor type, and numbers represent number of cases (percentages). Reverse target appearance was positive if the lesion showed initial peripheral enhancement and central hypoenhancement followed by delayed central enhancement or fill-in and decreased peripheral enhancement. Results Clinicopathologic Features The mean age, sex, tumor marker status, and risk factors for cirrhosis are shown in Table 1 for patients with combined hepatocellular cholangiocarcinoma tumors, cholangiocarcinoma, and HCC. Of the patients with tumors, 12 presented incidentally: nine with pain and three discovered on cirrhosis screening. Of those three, one presented with jaundice, one with a palpable mass, and one with consistently elevated liver function tests. In two patients, the presentation is unknown. Patients with combined had midrange elevation in both AFP and CA 19-9 compared with the control cases of cholangiocarcinoma and HCC patients. The average size of combined hepatocellular cholangiocarcinoma tumors was 7.7 ± 4.1 cm (median, 7.5 cm). The average size of intrahepatic cholangiocarcinoma tumors was 7.1 ± 2.8 cm (median, 6.85 cm). The average size of HCC tumors was 10.5 ± 5.5 cm (median, 9.0 cm). Of the patients with combined, 10.3% (n = 3) had evidence of lymphovascular invasion, one showed gallbladder invasion, and one showed involvement of the peribiliary glands. Metastatic disease was present in 17.2% of patients (n = 5), with two patients showing osseous metastases, one patient with lung metastases, and two patients with retroperitoneal lymph node metastases. Of patients who underwent hepatic resection, 10.3% (n = 3) had satellite lesions and 17.2% (n = 5) of specimens showed tumor necrosis. AJR:201, August

5 Fowler et al. In the 12 patients with combined hepatocellular cholangiocarcinoma tumors who were undergoing resection, two had undergone neoadjuvant transarterial chemoembolization (TACE), four had undergone adjuvant chemotherapy, and two had undergone combination radiation and chemotherapy. Patients who were not undergoing resection had been treated with a combination of chemotherapy and radiation, with three patients treated with TACE only. Of patients with combined hepatocellular cholangiocarcinoma tumors, six did not undergo further treatment at our institution after the histopathologic diagnosis. The average survival, based on time from initial imaging to death, in the 24 of 29 patients for whom the information was available was 15.5 months (median, 6 months; range, months). Imaging Findings The diagnosis of combined hepatocellular cholangiocarcinoma tumors on the basis of imaging features alone was suggested in 10 of 29 cases by reviewer 1 and in nine of 29 of cases A D Fig. 1 Combined hepatocellular cholangiocarcinoma in 57-year-old woman who presented with nausea and vomiting and was found to have large liver mass. A E, Unenhanced (A) and sequential contrast-enhanced MR images (B E) show large solitary T1 hypointense mass that peripherally and heterogeneously enhances during arterial phase followed by progressive delayed enhancement by 5-minute delayed images. F, Fat-suppressed T2-weighted image shows capsular retraction (arrow) and heterogeneous T2 signal intensity. On hepatic resection, this lesion was found to represent with predominant cholangiocarcinoma pathology. B E by reviewer 2. Of the combined hepatocellular cholangiocarcinoma tumor cases, 12 of 29 and 10 of 29 were misdiagnosed as cholangiocarcinoma whereas seven of 29 and nine of 29 were misdiagnosed as HCC by reviewers 1 and 2, respectively. HCC was correctly diagnosed in 12 of 15 and 14 of 15 cases and cholangiocarcinoma in 15 of 18 and 18 of 18 cases by reviewers 1 and 2, respectively. Neither reviewer misdiagnosed an HCC as a cholangiocarcinoma or vice versa. Reviewer 1 had sensitivity of 34% and specificity of 81% for diagnosis of combined. Reviewer 2 had sensitivity of 33% and specificity of 100% for diagnosis of combined hepatocellular cholangiocarcinoma tumors. Table 2 shows the individual imaging features of tumors, HCC, and cholangiocarcinoma. The combined hepatocellular cholangiocarcinoma tumors all were found to show arterial enhancement. The interobserver agreement for detection of washout in tumors was substantial (κ = 0.76), for delayed enhancement was moderate (κ = 0.43), for pseudocapsule and capsular retraction was slight (κ = 0.08 and 0.14, respectively), and for biliary ductal dilation and necrosis was fair (κ = 0.23 and 0.24, respectively). The reviewers also described whether the arterial enhancement was homogeneous, heterogeneous, or peripheral in nature. Table 3 shows the enhancement pattern for each tumor type. For tumors, the most common pattern of enhancement was peripheral, reported in 17 of 27 (63%) and 14 of 27 (52%) cases by reviewers 1 and 2, respectively. There was moderate agreement between observers in assigning enhancement pattern (κ = 0.44). Specific features seen only on MRI included T1 and T2 signal intensity and diffusion restriction. T2 hyperintensity was described in 90%, T1 hypointensity in 85%, and diffusion restriction in 94% of combined by both reviewers. Vascular invasion was identified by reviewers 1 and 2 in 13 of 29 and C F 336 AJR:201, August 2013

6 Imaging Biphenotypic Tumors 15 of 29 cases, respectively, with substantial agreement (κ = 0.72). Presence or absence of background liver cirrhosis, iron, and steatosis were determined by the reviewers. Loss of signal intensity on chemical-shift imaging on either in- or opposed-phase images, respectively, was used for determination of iron or fat deposition. Cirrhosis was defined as surface nodularity or evidence of portal hypertension. Both reviewers agreed that cirrhosis was present in six (20%), iron in one (3%), and fat in two (7%) of the 29 patients with combined hepatocellular cholangiocarcinoma tumors. Control cases were chosen to show a similar distribution of cirrhosis and steatosis, which was confirmed histologically in all cases. Discussion Combined hepatocellular cholangiocarcinoma tumors are a rare primary hepatic malignancy composed of a combination of mature and immature cholangiocyte-derived and hepatocyte-derived neoplastic elements. In A D Fig. 2 Combined hepatocellular cholangiocarcinoma in 84-year-old man with incidental liver lesion. A F, Arterial phase (A C) and portal venous phase (D F) CT images show large heterogeneously enhancing mass and areas of washout and subtle capsular retraction (arrow, D). Several satellite lesions are targetoid in appearance with peripheral enhancement (arrows, F). There is mild peripheral biliary ductal dilation in lateral segment (arrowhead, C). B E our retrospective study population, derived from 29 pathologically diagnosed cases of tumors, only a minority of patients had cirrhosis (20%) or positive hepatitis B or C serology (10%). This is similar to the findings of Jarnagin et al. [3], who published data on 27 patients, in whom hepatitis B or C positivity was seen in 15% and no patients had cirrhosis. However, this is in contradistinction to the majority of published demographic data, which suggest that cirrhosis and positive viral serology are common in patients with combined. To our knowledge, our study is only the second published regarding a U.S. population, and the discrepancy from previous literature may relate to different background population characteristics, such as the high rates of hepatitis B viral infection in Asian populations. The imaging features of combined hepatocellular cholangiocarcinoma tumors have not been the focus of most of the prior studies, and although it is commonly thought that the features overlap with both HCC and cholangiocarcinoma, the ability of radiologists to prospectively make the diagnosis and differentiate between these three entities has not previously been tested. Our study confirms that combined hepatocellular cholangiocarcinoma tumors may share some features with both cholangiocarcinoma and HCC but may resemble cholangiocarcinoma more closely in their enhancement pattern and ancillary features (Figs. 1 and 2). Radiologists were able to make the diagnosis of tumors only in a minority of cases but did show high specificity for the diagnosis when made. Differentiation of combined from cholangiocarcinoma proved most difficult and may not be possible prospectively. Perhaps more important to patient management is differentiation from pure HCC (Fig. 3). Although the imaging appearance C F AJR:201, August

7 Fowler et al. may overlap, some features might be useful in discriminating these lesions from HCC. In particular, the enhancement pattern, biliary ductal dilation, and capsular retraction should raise suspicion for cholangiocarcinoma or tumors (Fig. 2). Peripheral arterial enhancement was the most common pattern observed in our combined population but was also seen in the majority of cholangiocarcinoma cases as well. The targetlike or reverse targetlike pattern of enhancement, previously described by Aoki et A C Fig. 3 Combined hepatocellular cholangiocarcinoma in 49-year-old man who initially presented with jaundice and abdominal pain and was found to have large liver mass. A C, Arterial phase (A and C) and coronal portal venous phase (B) MR images show large heterogeneous mass with portal vein tumor thrombus expanding left and right portal veins. D, Equilibrium phase image shows washout within portions of tumor (arrow). Note curvilinear biliary drain in this patient who initially had peripheral ductal dilation and is currently decompressed. al. [29] and Ebied et al. [31] as being a common feature of tumors, was observed most frequently with cholangiocarcinoma in our study but was also seen in more than one third of tumors. Most notably, however, peripheral arterial and targetlike patterns of enhancement were rarely seen with HCC in our population. Knowledge of tumor markers may improve sensitivity for diagnosis. As suggested by previous studies, we found that combined showed substantially higher levels of CA 19-9 than HCC and potentially higher levels of AFP than cholangiocarcinoma. Hence, even in lesions with overlapping features, the diagnosis may be suggested on the basis of discordant imaging and tumor marker elevation. Our study has several limitations regarding the generalizability of the findings. First, the study is retrospective. Second, the patient population is relatively small, consisting of only 29 patients with tumors. Despite the small sample size, our population is similar in size to most other published series. Third, because of the retrospective nature of the study design, the im- B D 338 AJR:201, August 2013

8 Imaging Biphenotypic Tumors aging studies were heterogeneous on both CT and MRI. However, the relative rarity of this tumor type lends itself almost exclusively to retrospective studies, resulting in a mixture of possible imaging studies. Moreover, with the intent of defining imaging features on preoperative examinations, the appearance on both CT and MRI dynamic imaging is important because one or the other modality may be chosen for various reasons, such as contrast material allergy, renal dysfunction, or contraindications to MRI, such as pacemakers or implantable devices. Finally, it would have been ideal to analyze the tumor imaging features in conjunction with the composition of the lesion regarding relative cholangiocyte-hepatocyte derivation and maturity of the contained elements within the mixed tumors relative to the imaging phenotype. Fewer than one half of the cases underwent resection, and given the retrospective nature of the study, more detailed pathologic information was not available. Future studies may focus on this radiology-pathology correlation to better show how the relative composition of these mixed tumors translates into their imaging appearance. In conclusion, combined hepatocellular cholangiocarcinoma tumors can be challenging to diagnose on the basis of imaging features alone and appear to show morphology somewhere intermediate between that seen in HCC and cholangiocarcinoma. Enhancement pattern, ancillary imaging findings, and tumor markers may yield a relatively specific diagnosis in a minority of cases. Differentiation from cholangiocarcinoma may be particularly difficult. However, in cases in which the lesion does not show classic features of HCC, tissue sampling should be performed to confirm the diagnosis. References 1. Allen RA, Lisa JR. Combined liver cell and bile duct carcinoma. Am J Pathol 1949; 25: Theise N, Nakashima O, Park YN, Nakanuma Y. Combined hepatocellular-cholangiocarcinoma: WHO classification of tumours of the digestive system. Lyon, France: IARC, 2010: Jarnagin WR, Weber S, Tickoo SK, et al. Combined hepatocellular and cholangiocarcinoma: demographic, clinical and prognostic factors. Cancer 2002; 94: Yano Y, Yamamoto J, Kosuge T, et al. Combined hepatocellular and cholangiocarcinoma: a clinicopathologic study of 26 resected cases. Jpn J Clin Oncol 2003; 33: Zhang F, Chen XP, Zhang W, et al. Combined hepatocellular cholangiocarcinoma originating from hepatic progenitor cells: immunohistochemical and double-fluorescence immunostaining evidence. Histopathology 2008; 52: Itoyama M, Hata M, Yamanegi K, et al. Expression of both hepatocellular carcinoma and cholangiocarcinoma phenotypes in hepatocellular carcinoma and cholangiocarcinoma components in combined hepatocellular and cholangiocarcinoma. Med Mol Morphol 2012; 45: Woo HG, Lee JH, Yoon JH, et al. Identification of a cholangiocarcinoma-like gene expression trait in hepatocellular carcinoma. Cancer Res 2010; 70: Fujii H, Zhu XG, Matsumoto T, et al. Genetic classification of combined hepatocellular-cholangiocarcinoma. Hum Pathol 2000; 31: Yano H, Iemura A, Haramaki M, et al. A human combined hepatocellular and cholangiocarcinoma cell line (KMCH-2) that shows the features of hepatocellular carcinoma or cholangiocarcinoma under different growth conditions. J Hepatol 1996; 24: Kim H, Park C, Han KH, et al. Primary liver carcinoma of intermediate (hepatocyte-cholangiocyte) phenotype. J Hepatol 2004; 40: Tanaka S, Yamamoto T, Tanaka H, et al. Potentiality of combined hepatocellular and intrahepatic cholangiocellular carcinoma originating from a hepatic precursor cell: immunohistochemical evidence. Hepatol Res 2005; 32: Theise ND, Yao JL, Harada K, et al. Hepatic stem cell malignancies in adults: four cases. Histopathology 2003; 43: Seok JY, Na DC, Woo HG, et al. A fibrous stromal component in hepatocellular carcinoma reveals a cholangiocarcinoma-like gene expression trait and epithelial-mesenchymal transition. Hepatology 2012; 55: Yin X, Zhang BH, Qiu SJ, et al. Combined hepatocellular carcinoma and cholangiocarcinoma: clinical features, treatment modalities, and prognosis. Ann Surg Oncol 2012; 19: Koh KC, Lee H, Choi M, et al. Clinicopathologic features and prognosis of combined hepatocellular cholangiocarcinoma. Am J Surg 2005; 189: Liu CL, Fan ST, Lo CM, et al. Hepatic resection for combined hepatocellular and cholangiocarcinoma. Arch Surg 2003; 138: Ng IO, Shek TW, Nicholls J, Ma LT. Combined hepatocellular-cholangiocarcinoma: a clinicopathologic study. J Gastroenterol Hepatol 1998; 13: Chantajitr S, Wilasrusmee C, Lertsitichai P, Phromsopha N. Combined hepatocellular and cholangiocarcinoma: clinical features and prognostic study in a Thai population. J Hepatobiliary Pancreat Surg 2006; 13: Taguchi J, Nakashima O, Tanaka M, Hisaka T, Takazawa T, Kojiro M. A clinicopathological study on combined hepatocellular and cholangiocarcinoma. J Gastroenterol Hepatol 1996; 11: Kassahun WT, Hauss J. Management of combined hepatocellular and cholangiocarcinoma. Int J Clin Pract 2008; 62: Mathews J, Chapman WC, Vachharajani N, Tan B, Brunt EM. Primary liver carcinoma with biphenotypic expression: 23 cases of combined HCC- CCa. (abstr) Proceedings of the United States and Canadian Academy of Pathology annual meeting. Baltimore, MD: United States and Canadian Academy of Pathology 22. Wachtel MS, Zhang Y, Xu T, Chiriva-Internati M, Frezza E. Combined hepatocellular cholangiocarcinomas; analysis of a large database. Clin Med Pathol 2008; 1: Lee JH, Chung G, Yu S, et al. Long-term prognosis of combined hepatocellular and cholangiocarcinoma after curative resection comparison with hepatocellular carcinoma and cholangiocarcinoma. J Clin Gastroenterol 2011; 45: Yu XH, Xu LB, Zeng H, Zhang R, Want J, Liu C. Clinicopathological analysis of 14 patients with combined hepatocellular carcinoma and cholangiocarcinoma. Hepatobiliary Pancreat Dis Int 2011; 10: Zuo HQ, Yan LN, Zeng Y, et al. Clinicopathological characteristics of 15 patients with combined hepatocellular and cholangiocarcinoma. Hepatobiliary Pancreat Dis Int 2007; 6: Lee WS, Lee KW, Heo JS, et al. Comparison of combined hepatocellular and cholangiocarcinoma with hepatocellular and intrahepatic cholangiocarcinoma. Surg Today 2006; 36: Tang D, Nagano H, Nakamura M, et al. Clinical and pathological features of Allen s type C classification of resected combined hepatocellular carcinoma and cholangiocarcinoma: a comparative study with hepatocellular carcinoma and cholangiocellular carcinoma. J Gastrointest Surg 2006; 10: Sanada Y, Shiozaki S, Aoki H, Takakura N, Yoshida K, Yamaguchi Y. A clinical study of 11 cases of combined hepatocellular-cholangiocarcinoma: assessment of enhancement patterns on dynamic computed tomography before resection. Hepatol Res 2005; 32: Aoki K, Takayasu K, Kawano T, et al. Combined hepatocellular carcinoma and cholangiocarcinoma: clinical features and computed tomography findings. Hepatology 1993; 18: Takada K, Horita S, Meguro T, et al. A case of cholangiocellular carcinoma complicated with intratumoral hemorrhagic necrosis [in Japanese]. Nihon Shokakibyo Gakkai Zasshi 2004; 101: Ebied O, Federle MP, Blachar A, et al. Hepatocellular-cholangiocarcinoma: helical computed tomography findings in 30 patients. J Comput Assist Tomogr 2003; 27: AJR:201, August