Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Hepatocellular carcinoma (HCC) is the most common primary malignancy of the liver and accounts for 20% of all liver transplantation (LT) procedures performed in the United States.1 Mixed hepatocellular cholangiocarcinoma (HCC-CC) is a rare tumor with histological characteristics of both HCC and cholangiocarcinoma (CC).2, 3 Poor survival after HCC-CC resection has been reported,2, 4, 5 but data on outcomes after LT for HCC-CC are very limited.6, 7 CC tumors are divided into hilar/perihilar and intrahepatic subtypes. Intrahepatic cholangiocarcinoma (I-CC) is rare in patients with cirrhosis with etiologies other than primary sclerosing cholangitis (PSC).8 Successful outcomes have been reported after LT for select patients with hilar CC and PSC who have been treated with a neoadjuvant protocol,9 whereas I-CC is generally considered a contraindication to LT because of the high rate of posttransplant tumor recurrence.10-12
HCC-CC and I-CC were classified into 3 types by Allen and Lisa13 in 1949: type A for separate nodules of HCC and I-CC, type B for contiguous masses that may mingle, and type C for individual masses with HCC-CC. Another classification, which was formulated by Goodman et al.14 in 1985, includes type I for HCC and I-CC in separate nodules in the same liver, type II for HCC-CC, and type III for the fibrolamellar variant of HCC.
There have been reports describing the imaging features of HCC-CC,15, 16 which might be difficult to differentiate from HCC. Magnetic resonance imaging (MRI) features of I-CC also have been recently described in patients with cirrhosis before resection or LT.17 Because mixed HCC-CC and I-CC are relatively rare malignancies in patients with cirrhosis in comparison with HCC, mixed HCC-CC and I-CC may be underrecognized, misdiagnosed, and treated as if they were HCC in these patients. The fact that HCC and I-CC can coexist as separate nodules in the same liver (Goodman type I) makes the diagnosis even more challenging.
The primary aim of the present study was to evaluate the incidence, imaging features, and posttransplant outcomes of HCC-CC and I-CC discovered in the explant specimens of patients (excluding those with PSC) who underwent transplantation for HCC.
CC, cholangiocarcinoma; CK, cytokeratin; CT, computed tomography; HCC, hepatocellular carcinoma; HCC-CC, hepatocellular cholangiocarcinoma; I-CC, intrahepatic cholangiocarcinoma; LT, liver transplantation; MELD, Model for End-Stage Liver Disease; MRI, magnetic resonance imaging; PSC, primary sclerosing cholangitis; RFA, radio frequency ablation; TACE, transarterial chemoembolization; TNM, tumor-node-metastasis; UNOS, United Network for Organ Sharing.
PATIENTS AND METHODS
Between June 1999 and June 2009, 302 patients underwent LT for HCC in the setting of cirrhosis with etiologies other than PSC. According to a retrospective review of the LT database, 10 of these patients (3.3%) were found to have either HCC-CC or I-CC according to pathological examinations of explant specimens. Another 4 patients with incidental HCC-CC were identified among those patients without HCC who underwent transplantation for other indications during the same time period. The underlying liver disease in these 4 incidental cases was hepatitis C (3 patients) or autoimmune hepatitis (1 patient). The overall outcomes for this cohort of 14 patients with HCC-CC or I-CC were analyzed. With a matched-control study design (1:3), the outcomes and clinical and histological characteristics of the 10 patients (excluding the incidental cases) were compared with those of a control group of 30 transplant patients with only HCC in the explant; the study patients were matched with the control group by the number of lesions and the size of the largest nodule (±8 mm) in the explant. To minimize potential bias in the comparative analysis of the outcomes of the patients with HCC-CC or I-CC and the patients with a known diagnosis of HCC before LT, only the nonincidental cases were included in the matched-control study. This study was approved by the Institutional Review Board of the University of California, San Francisco.
All pretransplant imaging studies of the 10 patients with HCC-CC or I-CC and the 30 patients in the control group [including dynamic contrast-enhanced computed tomography (CT) or MRI with gadolinium contrast] were retrospectively reviewed by a radiologist experienced in the imaging of hepatobiliary malignancies so that he could identify radiographic features that might be different between the study and control groups. The analysis was focused on the contrast enhancement pattern of each lesion. The following scoring systems were used: (0) no enhancement, (1) mild enhancement, and (2) robust enhancement for arterial phase enhancement and (0) no washout, (1) mild washout, and (2) robust washout for venous phase washout. We also paid attention to the lesion size, the number of lesions, capsule retraction, and the location of hyperenhancement within lesions (peripheral or central).
For the analysis of tumor characteristics in liver explants, the tumor size, the number of lesions, the total tumor diameter, the pathological tumor stage (pT1-pT4) according to the modified United Network for Organ Sharing (UNOS) tumor-node-metastasis (TNM) classification,18 the grade of differentiation according to the Edmonson-Steiner criteria19 [(1) well differentiated, (2) moderately differentiated, and (3) poorly differentiated], and the presence or absence of vascular invasion were recorded. Explant tumor staging was based on the size and number of only viable tumors as previously reported.20
The histopathological definition of HCC-CC was based on the World Health Organization criteria.21 The diagnostic criteria included a hepatocellular element showing bile production, intercellular bile canaliculi, or a trabecular growth pattern and a cholangiocellular component showing mucin production or definitive gland formation. When the diagnosis was difficult, differences between I-CC and HCC were based on immunohistochemical studies that included cytokeratin 19 (CK19) or CK7 positivity, as previously reported.22, 23 The patients were classified according to the Goodman classification [type I (HCC and I-CC in separate nodules) or type II (HCC-CC)].14
Immunosuppressive Protocol and Postoperative Surveillance
The immunosuppression consisted of tacrolimus, mycophenolate mofetil, and a steroid (tapered to 5 mg of prednisone by posttransplant day 42). Rapamycin was not routinely used in patients undergoing LT for HCC. Six of the 14 patients of this particular cohort were given rapamycin after the discovery of HCC-CC or I-CC in the liver explant.
Data were reported as means and standard deviations or as medians and ranges when an abnormal distribution of data was identified. A nonparametric test (the Mann-Whitney U test) was used for numerical variables. The chi-square test with Fisher's correction was employed for categorical variables. Patient survival rates were estimated with the Kaplan-Meier method and were compared with the log-rank test. Statistical analyses were performed with SPSS 18.0 (SPSS, Inc., Chicago, IL). A P value less than 0.05 was considered statistically significant.
Baseline Clinical and Pathological Characteristics
The characteristics of all 14 patients with HCC-CC or I-CC, including the 4 patients with incidental HCC-CC, are summarized in Table 1. The majority of these patients had hepatitis C (57%) or hepatitis B (29%). All 10 nonincidental cases had undergone pretransplant locoregional therapy [most commonly transarterial chemoembolization (TACE)]. There were 4 patients with I-CC, 2 with Goodman type I tumors (HCC and I-CC), and 8 with Goodman type II tumors (HCC-CC). The last group of patients included the 4 incidental cases. CK19 staining was performed in 9 patients and was positive in 8 patients and negative in 1 patient. The tumor stage was T1 or T2 in the majority of the patients (71.4%). Poorly differentiated tumors were observed in 4 patients (30.7%), and microvascular invasion was found in 6 patients (43%). The patients with Goodman type II tumors (HCC-CC) had a higher incidence of vascular invasion (63% versus 17%, P = 0.08) and larger tumors (3.7 ± 1.5 versus 2.3 ± 2.1 cm, P = 0.1) than those patients with either I-CC or Goodman type I tumors (HCC and I-CC), but the differences were not statistically significant. No patients had evidence of lymph node invasion.
Table 1. Characteristics of Patients With I-CC or Mixed HCC-CC According to the Pathological Examination (n = 14)
The data are presented as medians and ranges.
The data are presented as means and standard deviations.
None of the 14 patients received preemptive adjuvant therapy after they were discovered to have I-CC or HCC-CC in the explant.
Tumor recurrence was observed in 57% of the patients (8/14). Six of the patients (75%) who had a recurrence had extrahepatic disease, 1 patient (12.5%) had an intrahepatic recurrence, and 1 patient (12.5%) had both intrahepatic and extrahepatic disease. Two of the 4 patients with incidental HCC-CC had a recurrence of their tumors. The median disease-free survival time for all 14 patients was 8 months (range = 1.25-84.8 months). The cumulative risks of recurrence at 1, 3, and 5 years were 40%, 50%, and 70%, respectively (Fig. 1). Four of the 8 patients who suffered from tumor recurrence (50%) received treatment consisting of a combination of chemotherapy and radiation therapy. The other 4 patients received only supportive care because of extensive extrahepatic metastases and their poor functional status. The median survival time after the diagnosis of tumor recurrence was 9 months (range = 0.4-23.4 months). No significant differences in disease-free survival were observed between incidental tumors [8 months (range = 2.8-45.4)] and nonincidental tumors [11.8 months (range = 1.25-84.8 months), P = 0.5]. Patients with I-CC or Goodman type I tumors (HCC and I-CC) had longer periods of disease-free survival [23.6 (range = 1.5-85 months) versus 8 months (range = 1.25-67.7 months), P = 0.4] and a better 5-year cumulative risk of recurrence (58% versus 81%, P = 0.5) in comparison with patients with Goodman type II tumors (HCC-CC), but the difference did not reach statistical significance. The 5-year cumulative risk of recurrence was 17% for HCC, 58% for HCC and I-CC or I-CC, and 78% for HCC-CC. These differences were statistically significant (P = 0.004; Fig. 2).
After a median follow-up period of 14 months (range = 1.8-68.8 months), the 1-, 3-, and 5-year survival rates after LT were 76%, 65%, and 51%, respectively (Fig. 3). Patients with either I-CC or Goodman type I tumors (HCC and I-CC) appeared to have longer actuarial survival times [34 months (range = 7.1-61.5 months) versus 10.3 months (range = 1.8-69 months), P = 0.5] than patients with Goodman type II tumors (HCC-CC), but the difference was not statistically significant.
Comparison of the Study and Control Groups
The characteristics of the study group of 10 patients (excluding the 4 incidental cases) and the control group of 30 patients with HCC are summarized in Table 2. Even though the 2 groups were matched according to the tumor size and number, the study group had a significantly lower proportion of patients with well-differentiated tumors (11.1% versus 43.3%, P < 0.02). There were no statistically significant differences in the incidence of microvascular or macrovascular invasion or any other characteristics (Table 2).
Table 2. Characteristics of the Study Group and a Control Group Matched by the Tumor Size and the Number of Nodules in the Explant
Study Group (n = 10)
Control Group (n = 30)
The data are presented as medians and ranges.
The data are presented as means and standard deviations.
Tumor recurrence was observed in 60% of the patients in the study group and in 16.7% of the patients of the control group (P = 0.008). The cumulative risks of recurrence at 1, 3, and 5 years were 42%, 54%, and 65%, respectively, in the study group and 10%, 17%, and 17%, respectively, in the control group (P < 0.002; Fig. 4).
The median patient survival time was 19.5 months (range = 1.8-67.7 months) in the study group and 34 months (range = 6.2-101 months) in the control group (P = 0.1). The differences between the study and control groups with respect to overall mortality (Table 2) and 1-, 3-, and 5-year actuarial survival (79%, 66%, and 47% versus 90%, 62%, and 62%, respectively, P = 0.1) did not reach statistical significance (Fig. 5). Nevertheless, the actuarial patient survival rates without recurrence at 1, 3, and 5 years were significantly higher in the study group versus the control group (90%, 62%, and 62% versus 79%, 45%, and 32%, respectively, P < 0.03; Fig. 6).
Preoperative Imaging Characteristics
The imaging characteristics of the tumors could be evaluated in 8 of the 10 patients in the study group. In the other 2 patients, the tumors were too small (<1 cm) to be fully characterized by preoperative imaging. All 8 patients had mild or robust arterial enhancement on CT and MRI. All imaging techniques showed a pattern of progressive contrast enhancement throughout the arterial and portal venous phases without washout (Fig. 7). This enhancement pattern was observed in both I-CC and HCC-CC and was seen prominently in the periphery of these lesions (Fig. 7). Two patients, one with HCC-CC and the other with both I-CC and HCC, had capsular retraction adjacent to the tumor. None of the tumors had bile duct dilatation according to imaging studies. All 30 patients in the control group, on the other hand, showed mild or robust arterial enhancement and washout in the delayed images.
Excellent 5-year patient survival rates exceeding 70% have been reported after LT for HCC.24, 25 Although successful outcomes after LT for the hilar and perihilar variants of CC have been reported for carefully selected patients treated with a neoadjuvant protocol,9 there is a general consensus that LT is contraindicated for patients with I-CC because of the high rate of tumor recurrence.10-12 HCC-CC accounts for 0.6% to 14% of all primary malignancies in the liver.2, 14 Both I-CC and HCC-CC have been described as aggressive tumors with poor survival rates and high recurrence rates after liver resection; the median survival period ranges from 20 to 47 months, and the 5-year recurrence rates are as high as 95% after resection.2, 4, 5 There are limited data on the incidence of HCC-CC and I-CC in patients undergoing transplantation for HCC because of a radiological misdiagnosis or an inability to differentiate these tumors from HCC by preoperative imaging; the data on posttransplant outcomes under these circumstances are also limited.
Here we report our experience with patients who underwent LT for HCC but were found to have either I-CC or HCC-CC during the explant pathological examination. Patients with PSC were excluded. We found an incidence of 3.3% for these tumors over a 10-year period. Chan et al.7 reported an overall incidence of 0.7% for HCC-CC according to explant pathological analysis among predominantly hepatitis B–infected patients who underwent LT. This incidence might have been higher if only patients who had undergone transplantation for HCC had been considered. Panjala et al.6 recently reported that among patients who underwent LT for HCC, 12 patients were found to have mixed HCC-CC and I-CC in the explant. The tumor recurrence rate was 58%, and the 5-year overall patient survival rate was a dismal 16%. We found a similar recurrence rate of 57% but a higher 5-year patient survival rate of 51%. In the study by Panjala et al., patients with Goodman type II tumors (HCC-CC) survived for a longer time than those with type I tumors (HCC and I-CC). In our analysis, we grouped together the patients with I-CC (without HCC) and those with HCC and I-CC (Goodman type I) under the assumption that their prognosis was mostly determined by the I-CC component. In contrast to the results of Panjala et al., we found that the patients with HCC-CC had a worse prognosis, a higher rate of tumor recurrence, and a shorter period of disease-free survival in comparison with the patients with I-CC. The small numbers of patients in the 2 studies preclude us from drawing firm conclusions about the outcomes of these tumor subtypes. In several published studies on liver resection, HCC-CC patients were found to have shorter periods of disease-free survival and poorer prognoses than I-CC patients2, 4, 26 and HCC patients.26 None of the patients in our study experienced lymph node invasion. The poor prognosis of patients with HCC-CC may be due to its inherent aggressiveness, which is linked to the origination of the tumor cells from pluripotent hepatic precursor cells.3, 14, 27
Our patients with HCC-CC and I-CC had worse posttransplant outcomes than the control group with HCC according to the cumulative risk of tumor recurrence and the recurrence-free survival rate. These patients had a 5-year cumulative recurrence rate of 65% and a 5-year recurrence-free survival rate of only 32%. These outcome data are consistent with the data reported for LT patients with PSC and I-CC, for whom a recurrence rate of 80%11 and a 5-year survival rate of 23%12 have been reported.
We evaluated the preoperative imaging characteristics of dynamic contrast-enhanced CT and/or MRI scans in all the patients, and we found that both I-CC and HCC-CC exhibited a pattern of progressive enhancement throughout the arterial and portal venous phases prominently in the periphery of the lesions. Contrast washout was not observed. These features clearly differed from the classic radiographic findings associated with HCC28 and were consistent with the descriptions in a recent report about I-CC in patients with cirrhosis before resection or LT.17 In addition, I-CC arising from a cirrhotic liver may be surrounded by a fibrotic pseudocapsule. As a result, capsular retraction may be seen along the tumor surface. This finding was observed in 2 patients in our series. It should also be noted that capsular retraction is usually not associated with HCC. None of our patients had biliary ductal dilatation adjacent to or surrounding the tumor, and thus the absence of biliary ductal dilatation should not exclude I-CC or HCC-CC from diagnostic consideration. In sharp contrast to our findings, Panjala et al.6 reported that the radiographic features of all their HCC-CC and I-CC cases were indistinguishable from those of their HCC cases. Panjala et al. did not independently reexamine all the imaging studies, and this could be one reason for the discrepancies. We performed a detailed assessment of all the preoperative cross-sectional imaging studies of our patients specifically to identify features distinguishing these tumors from HCC. Because HCC-CC may show radiographic characteristics of either HCC or I-CC, it is possible that only those tumors with features of I-CC were captured in the very small number of patients in the present series. In comparison, all patients in our control group with pathological confirmation of HCC had typical radiographic characteristics of HCC.
The present study has several limitations. The small number of patients with HCC-CC and I-CC and the retrospective study design may have led to bias related to the reported posttransplant outcomes and imaging characteristics of these tumors. It is also possible that some HCC-CC tumors and (less likely) some I-CC tumors have imaging features that are indistinguishable from those of HCC and thus cannot be used to exclude patients with these tumors from transplantation. The role of percutaneous biopsy for HCC before LT has been debated within the transplant community. Observing interval growth instead of biopsying small, indeterminate hepatic nodules has been advocated.28 HCC-CC and I-CC, on the other hand, represent contraindications to LT because of the very high rate of posttransplant tumor recurrence. It would be difficult, however, to exclude these patients from LT on the basis of suspicious imaging features alone without histological confirmation of the diagnosis by biopsy.
In summary, we have reported our experience with HCC-CC and I-CC, which were misdiagnosed and treated as HCC before LT in 3.3% of our patients over a 10-year period. These patients had a significantly higher rate of tumor recurrence and a shorter period of recurrence-free survival in comparison with a matched-control group of patients with only HCC in the explant. Both I-CC and HCC-C in our series exhibited imaging features that were distinct from those of HCC, and a pattern of progressive contrast enhancement throughout the arterial and portal venous phases was seen predominantly in the periphery of the tumor and without washout. A heightened awareness of HCC-CC and I-CC in patients with cirrhosis and the use of stringent imaging criteria for the diagnosis of HCC are important in order to prevent transplantation in patients with I-CC and HCC-CC. Patients with cirrhosis and tumors showing imaging features suspicious for I-CC and HCC-CC or imaging features not typical for HCC should undergo needle biopsy of the tumor. Once the diagnosis of HCC-CC or I-CC is confirmed histologically, these patients should be excluded from LT.