Annals of Surgical Oncology 15(7):1880–1890
DOI: 10.1245/s10434-008-9933-y
Intrahepatic Cholangiocarcinoma and Combined Hepatocellular-Cholangiocarcinoma: A Western Experience Nazario Portolani, MD,1 Gian Luca Baiocchi, MD,1 Arianna Coniglio, MD,1 Tullio Piardi, MD,1 Luigi Grazioli, MD,2 Anna Benetti, MD,3 Andrea Ferrari Bravo, MD,1 and Stefano Maria Giulini, MD1
1
Department of Medical and Surgical Sciences, Surgical Clinic, Brescia University, P.le Spedali Civili, 1, 25123 Brescia, Italy 2 Department of Radiology, Azienda Ospedaliera Spedali Civili di Brescia, Brescia, Italy 3 Department of Pathology, Brescia University, Brescia, Italy
Background: Intrahepatic cholangiocarcinoma (ICC) is an unusual tumour. Methods: The clinicopathological data of 67 patients with ICC and combined hepatocellular–cholangiocarcinoma (HCC–ICC) are presented. Results: HCV–HBV infection was present in 37.3% and chronic liver disease in 38.7% of cases, a rate higher than in the normal population; in these patients the cancer was small, often asymptomatic and of combined type. Liver resection was performed in 51 patients; at 1, 3 and 5 years, overall survival was 87.9%, 59.0%, and disease-free survival was 47.7% and 78.8%, 51.4%, and 46.7%, respectively. The better results were in the group of cirrhotic patients in whom ICC was diagnosed by a screening program for HCC (5-year survival 76.6%). Nodal metastasis showed negative prognostic value for both overall and disease-free survival; in N+ patients mean survival was 14.7 months after liver resection and lymph node dissection. Conclusion: Viral infection and cirrhosis may be considered risk conditions for ICC and combined HCC–ICC; in resected patients survival was good. Nodal metastases must not be considered a contraindication for liver resection. Key Words: HCC—Cholangiocarcinoma—Liver surgery—Hepatitis infection—Cirrhosis.
Intrahepatic cholangiocarcinoma (ICC) is the second most common primary liver tumour, but accounts for fewer than 10% of primary hepatic cancers.1 The paucity of observations limits knowledge of the pathogenesis of ICC and consequently identification of the patients at risk. This problem is particularly evident in Western countries where the factors generally linked to this cancer in Asian countries, such as parasitic infestation and diffuse intrahepatic lithiasis, are rarely observed,2 and no other risk factor has been reported. Consequently, diagnosis is generally late; when symptoms appear
the lesion is large and hepatic resection is rarely possible. Given the lack of specific aspects from computed tomography and magnetic resonance imaging, ICC is generally diagnosed only when the other more common primary and secondary hepatic malignancies are excluded, or it is assessed only at the post-resection pathological investigation, thus limiting the possibilities of a planned therapeutic strategy.3,4 Some adjunctive problems are related to a particular type of tumour in which the cholangiocarcinoma arises in combination with the hepatocellular carcinoma (HCC) in the same nodule or in the same liver. These forms, known as hepatocellular cholangiocarcinoma or combined HCC–ICC, are usually reported from Eastern surgical groups in small series and separately from ‘‘pure’’ ICC, suggesting a distinct pathogenesis
Published online April 29, 2008. Address correspondence and reprint requests to: Gian Luca Baiocchi, MD; E-mail:
[email protected] Published by Springer Science+Business Media, LLC Ó 2008 The Society of Surgical Oncology, Inc.
1880
1881
INTRAHEPATIC CHOLANGIOCARCINOMA
and clinical behaviour, leading to different prognosis after surgery.5–10 Consequently, no definite therapeutic approach is defined for these rare tumours. This study was planned with the aim of analysing the natural history of ICC and combined HCC–ICC and to discuss the problems currently encountered in their management, with particular focus on the pathogenic role of chronic liver diseases and the different factors influencing prognosis.
PATIENTS AND METHODS Between 1990 and 2006, 412 patients with primary liver cancer underwent hepatic resection at the Surgical Clinic of Brescia University, Italy; 51 patients (12.3%) were affected by cholangiocarcinoma. Patients with cancer arising from the main right or left hepatic bile ducts, the bifurcation, the extrahepatic biliary tree or the gallbladder, regardless of the presence of liver invasion, were excluded from this study. At pathological analysis, all the patients were found to have a neoplasm arising from the biliary epithelium, including pure ICC (33 patients, group A) and combined HCC–ICC (18 patients, group B). A further 16 patients with a pure ICC not submitted to liver resection were analysed, only with regard to risk factors and clinical data. Therefore, 67 patients are the object of this report, 49 belonging to group A and 18 to group B (Table 1). The most commonly employed diagnostic examinations were ultrasonography (US), computed tomography (CT) and magnetic resonance imaging (MRI), together with percutaneous needle biopsy (FNAB). FNAB was performed when preoperative imaging resulted atypical for HCC with suspicion of HCC or metastatic lesion. For each diagnostic technique, we evaluated the ability to detect the nodule, to assess malignancy, to suggest the nature of cancer (HCC, ICC, combined HCC–ICC, metastasis), and finally the accuracy in the preoperative staging, in relation to the presence of satellitosis, macroscopic vascular infiltration and nodal metastases. Hepatic resection was dictated by the oncological criteria of aiming to obtain a free margin of 0.5 cm whenever possible. In absence of contraindications such as cirrhosis with lymphatic congestion, when a preoperative diagnosis of ICC was available or suspected, a lymph node dissection was performed on the hepatic pedicle and common hepatic artery if the lesion was located in the right liver, extended to the celiac trunk and the left hepatic artery in case of left-located tumours.
TABLE 1. Clinicopathological characteristics of 67 patients presenting intrahepatic cholangiocarcinoma No. of patients Patients Mean age (range) Male/female Asymptomatic at diagnosis Liver status Cirrhosis Chronic hepatitis Viral markers HBV positivity HCV positivity Tumour markers Elevated aFP ([7 ng/ml) Elevated Ca 19.9 ([36 U/ml) Jaundice at diagnosis Tumour Pure ICC Combined HCC–ICC Mean size (range) Number of nodules 1 [1 Staging (AJCC) Stage I Stage II Stage III A (T3) C (N+) Stage IV Surgical treatment Resectability Type of surgery Major resections Minor resections Lymph node dissection Short-term survival Postoperative mortality Postoperative morbidity Major Minor
%
66 years (27–82) 36/31 36 53.7 17 9
25.3 13.4
8/67 17/65
11.9 26.1
23/60 29/52 6
38.3 55.7 8.9
49 73.1 18 26.8 5.8 cm (0.5–30) 50 17
74.6 25.3
18 8 6 13 6
35.4 15.6 11.7 25.4 11.7
51
76.1
28 23 34
54.9 45.0 66.6
3/51 8 13
5.8 15.6 25.5
All the pathologic specimens were evaluated by the same pathologist (A.B.). The tumour was staged according to the 2002 American Joint Committee on Cancer (AJCC) tumour–node–metastasis (TNM) classification;11 the parameters of the Liver Cancer Study Group of Japan were considered12 for the morphological classification of ICC. The lesions were studied to obtain histology (ICC or combined HCC– ICC), diameter, number of nodules, satellitosis, grading, vascular and neural infiltration and the width of the resection margin, as possible prognostic features. Whenever combined HCC–ICC was suspected, we combined the histological evaluation with the immunohistochemical tests for cytokeratins 7 and 19 expression (DBA production, diluted at 1:50, treated with three cycles of microwaves in EDTA), considered specific for biliary epithelium according to Nakamura;6 in some cases AE1, specific for keratin, was also Ann. Surg. Oncol. Vol. 15, No. 7, 2008
1882
N. PORTOLANI ET AL.
employed. For the hepatocellular component we used the monoclonal antibodies Hep Par 1 (DBA production, diluted at 1:50, treated with three cycles of microwaves in citrate); in some cases the antibodies for CD10 (synusoidal marker) and AE3 (keratin marker) were also used. Both histology and immunohistochemistry were considered necessary for conclusive diagnosis; these analyses were performed only on resected lesions. When a combined HCC–ICC tumour was diagnosed, a further classification was performed according to Goodman’s criteria:13 type 1 collision tumour, when separate and apparently coincidental areas of HCC and ICC arose in the same or in two separate nodules without any transition from one phenotype to the other, and type 2 transitional tumour, when an intimate intermingling of hepatocellular and glandular elements was present. In these combined HCC–ICC tumours extensive evaluation was performed to ascertain the relative components of the two tumours, confirming the diagnosis when a significant amount ([10%) of both HCC and ICC was present. Liver parenchyma adjacent to the tumour was studied to assess cirrhosis and chronic hepatitis, including both chronic persistent and active hepatitis; steatosis was considered together with mild nonspecific alterations and classified as ‘‘normal’’ liver. Long-term results were analyzed only for resected patients having more than 6 months follow-up; analysis was performed in the whole group of patients, separately in groups A and B, and finally compared to 213 patients submitted to liver resection for HCC in our unit in the same period. A univariate analysis of prognostic factors for overall and diseasefree survival was performed, considering sex, symptoms at presentation, presence of chronic liver disease, size of cancer, LN dissection, LN metastasis, vascular and neural infiltration. When recurrence occurred, we considered the free interval (early or late recurrence, using 2 years as the cut off) and the site (hepatic, nodal, peritoneal or systemic) and the feasibility of a potentially curative treatment. Cumulative survival rates were calculated by the Kaplan– Meier method and comparison between groups was performed by the Mantel–Haenszel test.
RESULTS Patient Characteristics The demographic data are reported in Table 1. Seventeen patients were cirrhotic (25.3%), all Ann. Surg. Oncol. Vol. 15, No. 7, 2008
belonging to Child class A, and nine patients were affected by chronic hepatitis (13.4%). Twenty-five patients (37.3%) presented one or more data related to an active viral infection (HCV antibodies and HBV surface antigens); this rate was 76.9% in the subgroup with cirrhosis or chronic hepatitis, compared to 12.1% in presence of a normal liver (P \ 0.01). The HCV-positivity rate was 18.1% in group A and 61.1% in group B patients (P \ 0.05). No patient had predisposing factors such as parasitic infection, primary sclerosing cholangitis, biliary atresia or previous thorotrast exposure; hepatolithiasis was found only in one case. Thirty-one patients (46.2%) were symptomatic, including abdominal pain in 19 patients, jaundice in 6, fever in 2 and weight loss in 6; a palpable mass was present in 4 patients. The tumour was found incidentally in 36 asymptomatic patients (53.7%); 9 of these were cirrhotic (one previously submitted to ethanol injection for HCC) and 1 was an HCVpositive patient with chronic hepatitis. These 10 patients were regularly followed in an HCC surveillance program including periodic ultrasonographic evaluations. Among the remaining asymptomatic patients 3 cases were ascertained during follow-up imposed by an initial diagnosis of haemangioma (mean diagnostic delay of 40 months), 6 cases during follow up evaluation for other tumours and 17 by US or CT scan performed for non-liver-related reasons. CA-19-9 was elevated above the normal value of 36 U/ml in 29 out of 52 patients (55.7%) but it was more than threefold in only 13 patients (25%), with relative positive values of 54.5% and 57.1% in group A and B patients, respectively; median value for positive cases was 234 (range 41–1289). The aFP value was high ([7 ng/ml) in 23 out of 60 cases (38.3%), with a positive rate of 10.3% in group A and 82.3% in group B (P \ 0.01); the value was above 400 ng/ml in 7 patients. Median value for positive cases was 49 (range 9–587). Diagnostic Reliability All the patients were submitted to US, 58 patients to CT, 27 to MRI and 6 to other examinations, including angiography (3 patients) and laparoscopy (3 patients); all the patients were submitted to at least two different examinations. At US the nodular lesion was detected in all but one case (98.5%) and its malignant nature was suggested in 55 out of 66 cases (83.3%), the remaining cases being evaluated as haemangioma (3 cases) or aspecific regeneration nodule (8 cases). Only in 7 cases (12.7%) was the
1883
INTRAHEPATIC CHOLANGIOCARCINOMA
TABLE 2. Diagnostic reliability of imaging techniques US (67 pts.)
Assessment of the nodule Diagnosis of malignancy° Right diagnosis of ICC or HCC–ICC^ Right staging§
CT scan (58 pts.)
MRI (27 pts.)
n
%
n
%
n
%
66 55* 7 35
98.5 83.3 12.7 63.6
56 46* 16 36
96.5 82.1 34.7 78.2
27 23* 11 17
100 85.1 47.8 73.9
° Defined as ‘‘suspected malignant nature’’ of the lesion, thus supporting a further evaluation, a percutaneous biopsy or directly the resection. ^ Defined as the presence of all the typical radiological features of this cancer, thus raising suspicion of ICC or combined HCC–ICC. § Defined as a correct evaluation of nodal metastases, macroscopic vascular infiltration and associated liver nodules, as confirmed by surgical exploration and/or pathological analysis. * Three patients after US, three after CT scan and one after MRI were considered as bearing an haemangioma.
lesion suspected to be a cholangiocarcinoma at US, the most frequent alternative diagnosis being HCC (18 cases). Correct staging was obtained only in 35 cases (63.6%); in the remaining cases, US missed the nodal metastases (13 patients), the macroscopic vascular infiltration (6 patients) and the associated nodular lesions (9 patients), as confirmed by surgical exploration and/or pathological examination. The corresponding values for CT were 96.5% (56 out of 58 lesions were detected), 82.1% (46 were considered malignant) and 34.7% (16 were rightly suspected as cholangiocellular type), while the correct preoperative staging was 78.2%; mistakes were linked to the nodal metastases (9 out of 13 were unknown), the macroscopic vascular infiltration (incorrect diagnosis in 27% of patients) and the associated nodules, not shown in 37.5% of patients. MRI was introduced in 1993 for studying this lesion; its efficacy was not statistically different from that of CT (Table 2). The final preoperative diagnosis, based on US followed by CT and/or MRI, was of ICC only in 26 patients (38.8%); the accuracy was significantly lower in cirrhotic patients (19.2%, P \ 0.05). The most frequent mistake concerned hepatic metastasis in group A (5 cases) and HCC in group B (11 cases). In six patients the lesion had been defined as benign (five atypical haemangiomas and one hepatic abscess), both at CT scan and MRI; the surgical indication arose from serial follow-up investigations showing a significant variation of the lesions. The initial mistake caused delay in diagnosis (from 24 to 60 months) and increase in size (medially 2 cm) in three patients, while in one patient the growth of satellite lesions was shown. A preoperative percutaneous biopsy was performed in 31 out of 51 resected cases; a reliable specimen was obtained in every case. The biopsy results are reported in Table 3: correct ICC diagnosis was obtained in 45.4% of group A patients, while the
TABLE 3. Histology after percutaneous biopsy (FNAB)* Final pathologic diagnosis Combined HCC–ICC (9 pts.)
ICC (22 pts.) N Diagnosis after percutaneous biopsy ICC 10 Combined HCC–ICC 0 HCC 1 Metastases 4 Undefined diagnosis** 7
%
n
45.4
1 1 3 4 0
%
11.1
* Data regarding 31 resected patients in which a percutaneous biopsy was made. ** No distinction was possible between ICC and metastasis (five cases), HCC (one case), and between HCC and metastasis (one case).
histological evaluation suggested combined HCC– ICC only in 11.1% of group B patients. Surgical Data and Pathological Analysis of the Resected Specimen All the patients were submitted to surgical exploration. Fifteen patients, operated on for a large cancer (mean size 11.5 cm), were being resected owing to diffuse peritoneal involvement or a nodal or hepatic extension greater than preoperatively suspected. An 82-year-old patient in poor general condition underwent intraoperative radiofrequency ablation of a deep nodule, 3.5 cm in size, located between segments V and VIII; this decision was taken when the liver proved to be cirrhotic at laparotomy. The remaining 51 patients were subjected to resection. The resectability rate was 100% in the subgroup of patients with chronic liver disease submitted to a screening program for liver cancer, 85.7% and 62.1%, respectively in the asymptomatic and symptomatic patient subgroups; in patients with Ann. Surg. Oncol. Vol. 15, No. 7, 2008
1884
N. PORTOLANI ET AL.
TABLE 4. Pure ICC and combined HCC–ICC in patients with liver cirrhosis (LC) or chronic hepatitis (CH) compared to the patients with normal liver (NL) LC + CH (26 patients) n Asymptomatic discovery Rate of HBV infection Rate of HCV infection Size of the lesion (mean ± SD) Preoperative diagnosis of ICC Reliability of percutaneous biopsy Rate of combined HCC–ICC* Resection rate Rate of major resection* Rate of lymph node dissection*
15/26 5/26 15/25 4.6 (±2.41) 5/26 5/15 14/25 25/26 10/25 12/25
% 57.6 19.2 60.0 19.2 33.3 56.0 96.1 40.0 48.0
NL patients (41 patients) n 21/41 3/41 2/40 8.0 (± 4.05) 22/41 18/28 4/26 26/41 18/26 22/26
%
P
51.2 7.3 5.0
N.S. N.S. \0.01 \0.05 \0.05 \0.05 \0.01 \0.05 \0.05 \0.05
53.6 64.2 15.3 63.4 69.2 84.6
* Values are related to 51 resected patients.
chronic liver disease resection was possible in 96.1% of cases (Table 4). Resectability rate was not influenced by the value of preoperative tumour markers (P = 0.24). We performed 28 major resections (54.9%), 14 mono- or bisegmental resections (27.4%) and finally 9 subsegmentectomies (17.6%), confined to the group of cirrhotic patients. Resection included the extrahepatic biliary tree and the caudate lobe in six patients. In two patients an extended right hepatectomy and a right hepatectomy were, respectively, associated with resection of the retrohepatic vena cava, portal confluence, right adrenal gland and peritoneum in one patient, and with splenectomy, omentectomy and peritonectomy in the other. Lymph node dissection was performed in 34 patients in which a preoperative diagnosis of ICC was available (66.6%), with a significant difference relating to the presence (48.0%) and absence (84.6%) of chronic liver disease (P \ 0.05, Table 4), and to the extension of hepatic resection (76% in major and 52.3% in minor hepatectomies). The remaining 17 patients were submitted only to hepatic resection; in 11 patients because the diagnosis of ICC was obtained only postoperatively, while in six cirrhotic patients with a small ICC (mean size 4 cm) without any clinical and radiological sign of nodal involvement, the presence of severe portal and lymphatic hypertension led to a more conservative approach. At pathological study, 81 lesions were analyzed, with 15 cancers resulting as multinodular. At macroscopy, the lesion was hard in consistency with diffuse fibrosis in 22 patients (43.1%). The mean size was 5.8 cm (range 0.5–30 cm); a significant difference was noted comparing the patients affected by cirrhosis or chronic hepatitis (4.6 cm) with the normal liver patients (8.0 cm, P \ 0.05) and the symptomatic with the asymptomatic patients (9.3 versus 5.1 cm, P \ 0.01). The asymptomatic cirrhotic patient subAnn. Surg. Oncol. Vol. 15, No. 7, 2008
group screened for HCC showed the smallest size (mean 3.4 cm), less than 5 cm in 90% of cases. In 33 cases (64.7%, group A) the final diagnosis was pure cholangiocarcinoma. The tumours were classified according to macroscopic appearance as mass-forming type (type 1) in 30 cases, periductal infiltrating lesions, i.e. tumour infiltration along the bile duct (type 2) in 2 cases, while in 1 case a type 4 lesion was observed, i.e. a mass-type lesion showing an infiltration along the hilar plane. Eighteen patients (35.2%, group B) had a combined HCC–ICC tumour; 14 had chronic hepatitis or cirrhosis, and 11 were HCV positive (Table 4). Four combined tumours were considered as collision type (including three patients with two separate nodules) and 14 as transitional type. In these last lesions, the prevalent component was HCC in five cases (80%, 90%, 90%, 90% and 90%), ICC in four cases (60%, 90%, 90% and 90%); in the remaining five cases the extensive examination of the nodule showed a similar incidence of the two components (50%:50%). Vascular and neural infiltration were documented with a similar rate in groups A (49.3%) and B (41.7%). A total of 334 lymph nodes were removed from 34 patients (range 2–33). Nodal metastases were diagnosed in 13 patients, with the same incidence in groups A and B (Table 5). A significant trend was observed deriving from cancer size: the rate of nodal metastases was 27% for cancer measuring less than 5 cm, 44.6% when the size was between 5 and 10 cm and finally 100% for nodules larger than 10 cm. Early and Long-Term Survival and Recurrence Analysis Postoperative mortality was 5.8%: three cirrhotic patients died, two from liver failure and one from
1885
INTRAHEPATIC CHOLANGIOCARCINOMA
TABLE 5. Characteristics of patients with pure ICC (group A) and combined HCC–ICC (group B)* Group A (33 patients)
Asymptomatic discovery Rate of HBV infection Rate of HCV infection Elevated CA 19.9 Elevated aFP Rate of cirrhosis/chronic hepatitis Mean size at diagnosis (cm) Rate of vascular/neural infiltration Rate of nodal metastases
Group B (18 patients)
n
%
n
%
P
19/33 5/33 6/33 12/22 3/29 11/33 5.8 21/33 9/25
57.5 15.1 18.1 54.5 10.3 33.3 7.0 63.6 36.0
7/18 3/18 11/18 8/14 57.1 14/17 14/18 N.S. 13/18 4/9
38.8 16.6 61.1 N.S. 82.3 77.7
N.S. N.S. \0.05 \0.01 \0.01
72.2 44.4
N.S. N.S.
* Data are related to 51 resected patients.
1,0
Overall Survival (%)
sepsis. Major postoperative morbidity was 15.6%; two patients required reintervention for haemoperitoneum, two patients underwent percutaneous drainage of abdominal collections after biliary leakage and four patients suffered liver failure. The minor morbidity rate was 25.5%, mainly due to pleural effusion and prolonged postoperative ascites. In cirrhotic patients submitted to liver resection and lymph node dissection, prolonged postoperative ascites was documented in 60% of cases. After surgery, only three patients were treated with adjuvant chemotherapy. No patient was lost to follow-up (mean 28.5 months, range 8–163 months). Mean survival of the patients submitted only to explorative laparotomy was 4 months. After resection, mean survival was 37.8 months; 20 patients died, recurrence being the main cause of death (18 patients), followed by chronic liver failure (2 patients). The disease-specific survival rate was 87.9%, 59.0% and 47.7% at 1, 3 and 5 years, without a significant difference between group A (86.4%, 56.1%, 42.9%) and B (90.9%, 62.3%, 62.3%) (Fig. 1). These data are similar to the large group of 213 patients submitted to liver resection for HCC in the same period (87.5%, 61.1% and 41.9%). Recurrence developed in 18 patients (35.2%), all but two in the first 2 years. Disease-free survival is reported in Fig. 2; no difference was noted between group A and group B. The ten cirrhotic patients with asymptomatic lesions discovered during routine follow-up had the best survival results compared to the remaining patients, both overall (89.4%, 76.6% and 76.6% versus 81.4%, 49.2% and 27.3% at 1, 3 and 5 years) and for disease-free survival (89.4%, 63.9% and 63.9% versus 62.9%, 34.1% and 34.1% at 1, 3 and 5 years); these differences were statistically significant (P \ 0.05). Recurrence sites included the liver (13 cases), the retroperitoneal or hilar nodes (2 cases), the chest (2
Cumulative Pure ICC (Group A) Combined ICC-HCC (Group B)
0,8
0,6
0,4
0,2 0
20
40
60
80
100
Time (m) FIG. 1. Overall survival after liver resection for intrahepatic cholangiocarcinoma in the whole group (51 patients), in pure ICC (33 patients) and in combined ICC–HCC (18 patients).
cases), the peritoneum (2 cases), the adrenal (1 case), the brain (1 case), the abdominal wall (1 case) and bone (1 case); in 9 cases (50%) the liver was the only recurrence site. Recurrence diagnosis was performed by CT scan in every case; percutaneous biopsy was never necessary. Four patients underwent surgical exploration after recurrence and in two a curative iterative resection was possible. Five patients underwent chemotherapy and one was treated by external bone radiotherapy, while only symptomatic medical therapy was administered in the remaining eight patients. Overall, mean survival after recurrence was 9.4 months. The two patients submitted to redo surgery died after 7 and 17 months, while the mean survival after chemotherapy was 13 months; no patient treated with palliative medical intent survived for more than 4 months. Ann. Surg. Oncol. Vol. 15, No. 7, 2008
1886
N. PORTOLANI ET AL.
1,0
1,0
NN+
0,8 0,8
Survival rate (%)
Disease free Survival (%)
0,9
0,7 0,6 0,5
0,6
0,4
0,2 0,4 0,3
0,0 0
20
40
60
80
100
0
10
FIG. 2. Disease-free survival in 51 resected patients.
20
30
40
50
60
Time (m)
Time (m)
FIG. 3. Overall survival in patients with (13 patients) and without (21 patients) nodal metastases undergoing lymphadenectomy.
Prognostic Factors Only the nodal metastasis (N+) was predictive of poor survival after surgery, while sex, symptoms on presentation, chronic liver disease, cancer size, LN dissection and vascular/neural infiltration were not. In N+ patients, mean survival was 14.7 months and only one patient survived longer than 36 months; the survival rate was 57.1%, 14.3% and 14.3% at 1, 3 and 5 years. On the contrary in N– patients mean survival was 38.5 months, with a survival rate of 84.6%, 76.2% and 76.2% at 1, 3 and 5 years (P \ 0.01, Fig. 3). The prognostic value of nodal metastasis was observed in both group A and B. No other factor related to the neoplasm and the type of surgery afforded prognostic value; lymph node dissection was not associated with longer survival. Two features were predictive of a high recurrence rate: vascular and neural infiltration (P \ 0.05) and nodal metastases (P \ 0.05, Fig. 4). The recurrence rate was 30% in the group of screened cirrhotic patients and 53.1% in the remaining patients (P = 0.2).
DISCUSSION Intrahepatic cholangiocarcinoma includes various types of cancer. On analysis, two aspects are of particular interest: combined lesions, i.e. hepatocellular carcinoma–cholangiocarcinoma (HCC–ICC), accounting for 26.8% of cases in our experience, and the presence of cirrhosis–chronic hepatitis in the surrounding liver, diagnosed in 38.7% of our patients. Ann. Surg. Oncol. Vol. 15, No. 7, 2008
Combined HCC–ICC Several different hypotheses have been suggested regarding the histogenesis of this unusual tumour: (1) two totally different and independent cancers coexist; (2) the cancer develops from a cellular line, generally the hepatocytes, and later transforms into the second component with transitional aspects; (3) the cancer develops from a liver stem cell, potentially evolutive in the two different cellular lines.14 Our experience suggests that no single hypothesis may explain all the observations; the group of combined tumours seems to be heterogeneous. In 1998 Nagafuchi reported two cases of resected double cancers, confirming the first hypothesis.15 In three cases we observed two completely distinct cancers and in a fourth patient a collision type of cancer without any transitional aspect within. In these cases the lack of histological transition and the absence of any contact between two separate different nodules with clear different immunohistochemical staining suggest the different nature of the neoplastic cells and support a distinct histogenesis of the tumours. In agreement with the second and the third hypotheses, intermediate histological features were a prominent aspect in the remaining 14 cancers. For these cancers the use of a specific marker for the progenitor cell, as recently proposed by Kim et al.,5 might further support the potential role of the stem cells especially when the transitional areas are highly represented in the lesion, as observed in 71.4% of our cases. In these tumours, better considered as mixed than combined HCC– ICC, the relative rates of the two cancers were widely
INTRAHEPATIC CHOLANGIOCARCINOMA
different, with a large prevalence also of the biliary component, as in four of our cases, confirming a similarity to ICC for these lesions, as recently suggested by genetic studies.10 On the contrary, some authors have stressed characteristics related both to biological and clinical behaviour for these combined tumours, making them more like HCC than ICC.9,16–18 However, comparison is generally made between numerically widely different groups, thus limiting the value of statistical comparison. Also, no consideration is afforded to the respective weight of the two different cellular lines within the combined lesion, giving the prevalent ICC or HCC aspect to the lesion in different cases. We observed multiple combinations; in the four cases with nodal metastases, the ‘‘metastatic cellular line’’ was hepatocellular in one case and cholangiocellular in three. Cancer aggressiveness features, such as vascular infiltration and nodal metastases, were similar in pure ICC and combined HCC–ICC, but more frequent than in HCC; the post-resection prognosis was also similar in groups A and B, but also not significantly different from the values observed in HCC patients.19 The Role of Viral Infection and Cirrhosis Our experience shows a significant incidence of chronic viral infection and liver cirrhosis associated to ICC. Above all, in combined ICC–HCC, the rate of HCV antibody positivity (61.1%), as well as the presence of cirrhosis or chronic hepatitis (77.7%) are comparable to values generally reported in the literature20,21 and to the personal experience of HCC.19 Also for pure ICC, generally considered arising in a normal liver, these rates are remarkable; 33.3% of patients had a chronic liver disease and 33.3% had a viral infection. Although this study was not built as a population study, these data — significantly higher than HCV and HBV positivity in the adult population living in the province of Brescia (4% and 5.2%,21 respectively) — are among the highest reported in the literature17,22 and have recently found some confirmatory reports in Eastern literature. In 2002 Okuda reported the results of a collective review of 982 ICC cases among Japanese patients; HCV antibodies were positive in 27.7%, 16.8% of the patients had chronic hepatitis and 10.9% were cirrhotic.2 In a prospective study on 600 patients with HCV-related cirrhosis, Kobayashi observed 14 ICCs in a period ranging from 0 to 18.5 years (median 7.2 years) of follow-up with an incidence of 0.3% per year; this rate was about 1,000-fold higher than the estimated incidence
1887
of ICC in the general population in Japan.22 In two recent Eastern ICC surveys, HBV infection was documented, respectively, in 31.8% and 36.2% of the patients and cirrhosis in 18.2% and 46.8%.17,23 These correlations are often lacking in the few reports from Western countries; when assessed, the HCV-positivity rate is significantly lower, between 3.7% and 12%,3,24 and a pathologic finding of cirrhosis is unusual, generally having an incidence lower than 6%.3,4,25 These differences are not easily explained, but can be linked to the different incidence of viral infection and of cholangiocarcinoma in the different countries;26 as well as for the most common HCC, the etiologic role of every single factor is widely different according to the geographic situation. We operate in a high-incidence area both for chronic viral infection and HCC;20 this study suggests how this factor may have a relevant etiologic role for ICC too. Therefore, the subclinical HCC screening program in cirrhotic patients may also help to discover some small cholangiocarcinomas, both in pure and combined form, as in our experience: nine out of ten lesions diagnosed in screened patients were smaller than 5 cm, resection rate was 100% and seven of these patients are disease free, a median of 39 months after surgery. The presence of a chronic liver disease gives some specific features to cholangiocarcinoma: the tumour is often combined, diagnosis is generally made early and in an asymptomatic stage, curative resection is often possible and the prognosis is relatively favourable after resection. Diagnostic Considerations The difficulty of obtaining a correct preoperative diagnosis of cholangiocarcinoma is underlined by several authors,3,4,24,25,27 but generally not discussed in detail. An initial possible mistake is identifying the lesion as a benign tumour. We observed six cases initially considered as benign, not only after US (three patients), but also after CT (four patients) or MRI (two patients). The most frequent misdiagnosis was ‘‘atypical haemangioma’’, as in five of our patients. When an atypical benign lesion is diagnosed, an accurate follow-up must be planned, especially in an area with a high incidence of hepatic cancer; diagnostic delay may compromise chance of survival. Problems increase when the requirement is to differentiate cholangiocarcinoma from HCC. This difficulty is particularly relevant for combined tumours and especially for cancers arising in a cirrhotic liver. Sasaki described seven cases of combined tumours; all these cancers were diagnosed as HCC.28 In our Ann. Surg. Oncol. Vol. 15, No. 7, 2008
1888
N. PORTOLANI ET AL.
1,0
No vascular/neural infiltration Vascular/neural infiltration
Disease free Survival (%)
0,9 0,8 0,7 0,6 0,5
Therapy, Prognosis and Survival
0,4 0,3 0
20
40
60
80
100
Time (m) 1,0
Disease free Survival (%)
macroscopic evaluation, may be considered a useful guide for surgical strategy. On the contrary, for a cancer in normal liver, the differential diagnosis is mainly between primary neoplasm and metastasis. Correct discrimination is not always obtained even with the aid of percutaneous biopsy, which can fail to differentiate ICC from metastasis. In these cases an extensive study in search of an eventual primary cancer is recommended.
NN+
0,8
0,6
0,4
0,2
0,0 0
20
40
60
80
100
Time (m) FIG. 4. Disease-free survival according to: (A) vascular and/or perineural infiltration (34 patients were positive for vascular/neural infiltration and 17 were negative) and (B) nodal metastases (among patients undergoing lymphadenectomy, 13 were N+ and 21 were N–).
experience, 9 out of 18 combined ICC–HCCs were diagnosed as HCC as well as 14 out of 26 patients with ICC arising in chronic liver disease. Misinterpretation of the preoperative evaluation has a therapeutic value. HCC is usually treated by surgery, but also with ethanol injection (PEI) and radiofrequency ablation (RFA), methods which are not widely tested for ICC; at surgery, in the case of HCC, just liver resection is required, owing to the low incidence of nodal metastases, while in presence of ICC the indication of lymph node dissection, as later discussed, is suggested by several authors. According to this consideration, wider use of intraoperative frozen biopsy, especially for the lesions not typical for HCC at Ann. Surg. Oncol. Vol. 15, No. 7, 2008
Liver resection offers a substantial chance of cure to patients with ICC, especially when early diagnosis is performed.29 The incidence of nodal metastases, vascular and neural infiltration, which are the best determinants of disease-free survival, increases with the size of the cancer. In particular, the nodal metastases are the main negative prognostic factor for the survival; nevertheless, there is no consensus about the indication to perform lymph node dissection in association to liver resection.25,28,30 No randomised trial focussed on this problem has been performed;31,32 our study too, based on a retrospective evaluation, cannot make a relevant contribution to this topic. In our opinion, the suspicion of nodal metastases should not be used to deny surgery; the preoperative diagnosis is difficult and false-positive reports are possible, especially with HBV–HCV infection, frequently associated to large immunoreactive nodes in the hepatic pedicle.33 Even presence of proven nodal metastases should not per se justify abstention from the planned operation. Long-term survival of N+ patients cannot be excluded according to some surgical experiences reporting small series of patients.9,18,34,35 In our N+ patients undergoing resection, survival is not negligible (mean survival 14.7 months) and one patient is alive more than 3 years after surgery. The reports of Ohtsuka, Koh and Nozaki,1,17,30 who found similar survival after lymph node dissection in patients with or without nodal metastasis, tend to further support the indication for liver resection with lymph node dissection in these cases. While awaiting prospective evaluation of large series of patients, the indication for lymph node dissection associated to liver resection for ICC cannot be strongly recommended, but must be considered, at least for a staging purpose.36 In our experience, considering the hypothesis of a therapeutic advantage, the best candidates for lymph node dissection are patients with large cancers, statistically associated with nodal metastases. A more conservative attitude may be considered in cirrhotic patients, in whom
1889
INTRAHEPATIC CHOLANGIOCARCINOMA
hepatic pedicle dissection may increase postoperative morbidity,33 especially if the lesion is of small diameter, which is associated with low risk of nodal involvement. The final consideration concerns cancer recurrence; the short delay (in about 90% of cases the recurrence is early) and the correlations with vascular and neural infiltration in the first lesion support the opinion that these recurrent lesions are real metastases instead of new cancers. Unlike HCC, in which recurrence is exclusively intrahepatic in more than 85% of the cases,16 the liver is the only failure site in no more than 50% of cases. The contemporary involvement of multiple sites and multicentric liver involvement usually limit the chance of treatment.1,17 Therefore, differently from HCC, recurrence generally indicates the definitive failure of the therapeutic program; the lack of an effective therapy for recurrence, more than intrinsic biologic aggressiveness, may be the real reason why the data points to poor ICC survival compared to HCC.16 Nevertheless, an iterative treatment based on RFA,37 or, in a very selected group of patients, on redo surgery eventually supported by complementary chemotherapy,36,38 cannot be excluded a priori if technically feasible and at low risk; survival seems to be higher with this therapeutic approach than after supportive care.
CONCLUSIONS AND RECOMMENDATIONS Intrahepatic cholangiocarcinoma may be surgically treated without discouraging results; the rate of longterm survival may be similar to the results of surgery for HCC. Pure ICC and combined ICC–HCC do not present significant biological differences and can be considered as a single entity from a clinical point of view. The frequent presence of a chronic liver disease selects a group of patients at risk for primary liver tumours, including ICC, and favours early diagnosis if the patients are regularly followed up. In the normal liver, cholangiocarcinoma may mimic a benign tumour, above all haemangioma: particularly when some typical aspects of this benign lesion are absent, a confirmatory test and a surveillance program are warranted, especially in geographic areas with a high incidence of primary liver cancers. Small lesions diagnosed from the screening program and free from nodal metastases offer a good chance of long-term survival when resected (75% at 5 years in our experience). Also, for larger tumours, surgery may allow a chance for long-term survival. Particularly in these patients, lymph node dissection can be envisaged, at
least for staging purposes, if we consider that in these advanced stages the adjunct value of chemotherapy must not be neglected, according to some recent experience of combined therapy36,38 and our encouraging results in the presence of liver recurrence. The value of this aggressive combined therapy (resection and chemotherapy) should be further investigated.
REFERENCES 1. Ohtsuka M, Ito H, Kimura F, et al. Results of surgical treatment for intrahepatic cholangiocarcinoma and clinicopathological factors influencing survival. Br J Surg 2002; 89:1525–31. 2. Okuda K, Nakanuma Y, Miyazaki M. Cholangiocarcinoma: Recent progress. Part 1: Epidemiology and etiology. J Gastroenterol Hepatol 2002; 17:1049–55. 3. Weber SM, Jarnagin WR, Klimstra D, et al. Intrahepatic cholangiocarcinoma: resectability, recurrence pattern, and outcomes. J Am Coll Surg 2001; 193:134–41. 4. Huang JL, Biehl TR, Lee FT, et al. Outcomes after resection of cholangiocellular carcinoma. Am J Surg 2004; 87:612–7. 5. Kim H, Park C, Han KH, et al. Primary liver carcinoma of intermediate (hepatocyte-cholangiocyte) phenotype. J Hepatol 2004; 40:298–304. 6. Nakamura S, Suzuki S, Sakaguchi T, et al. Surgical treatment of patients with mixed hepatocellular carcinoma and cholangiocarcinoma. Cancer 1996; 78:1671–6. 7. Ng IO, Shek TW, Nicholls J, et al. Combined hepatocellularcholangiocarcinoma: a clinicopathological study. J Gastroenterol Hepatol 1998; 13:34–40. 8. Taguchi J, Nakashima O, Tanaka M, et al. A clinicopathological study on combined hepatocellular and cholangiocarcinoma. J Gastroenterol Hepatol 1996; 11:758–64. 9. Jarnagin WR, Weber S, Tickoo SK, et al. Combined hepatocellular and cholangiocarcinoma: demographic, clinical, and prognostic factors. Cancer 2002; 94:2040–6. 10. Cazals-Hatem D, Rebouissou S, Bioulac-Sage P, et al. Clinical and molecular analysis of combined hepatocellular-cholangiocarcinomas. J Hepatol 2004; 41:292–8. 11. Green FL (ed.) AJCC cancer staging manual, 6th edn. New York: Springer Verlag, 2002. 12. Liver cancer study group of Japan. Classification of primary liver cancer, 1st edn. Tokyo: Kanahara, 1997. 13. Goodman ZD, Ishak KG, Langloss JM, et al. Combined hepatocellular-cholangiocarcinoma. A histologic and immunohistochemical study. Cancer 1985; 55:124–35. 14. Kojiro M. (1992) Patomorphology of advanced hepatocellular carcinoma. In: Tobe K, Kameda H, Okudaira Met al. (eds). Primary liver cancer in Japan. Springer-Verlag, Tokyo pp 31–37. 15. Nagafuchi Y, Okamoto K, Shono M, et al. Separate histogenesis of combined hepatocellular and cholangiocellular carcinoma in two patients. Hepatogastroenterology 1998; 45: 523–7. 16. Liu CL, Fan ST, Lo CM, et al. Hepatic resection for combined hepatocellular and cholangiocarcinoma. Arch Surg 2003; 138:86–90. 17. Koh KC, Lee H, Choi MS, et al. Clinicopathological features and prognosis of combined hepatocellular cholangiocarcinoma. Am J Surg 2005; 189:120–5. 18. Yano Y, Yamamoto J, Kosuge T, et al. Combined hepatocellular and cholangiocarcinoma: a clinicopathologic study of 26 resected cases. Jpn J Clin Oncol 2003; 33:283–7.
Ann. Surg. Oncol. Vol. 15, No. 7, 2008
1890
N. PORTOLANI ET AL.
19. Portolani N, Coniglio A, Ghidoni S, et al. Early and late recurrence after liver resection for hepatocellular carcinoma (HCC). Prognostic and therapeutic implications. Ann Surg 2006; 243:229–35. 20. Chiesa R, Donato F, Portolani N, et al. Primary liver cancer in a high incidence area in north Italy: ethiological hypotheses arising from routinely collected data. Eur J Epidem 1995; 6:435–42. 21. Donato F, Tagger A, Gelatti U, et al. Alcohol and hepatocellular carcinoma : the effect of lifetime intake and hepatitis virus infections in men and women. Am J Epidemiol 2002; 155:323–31. 22. Kobayashi M, Ikeda K, Saitoh S, et al. Incidence of primary cholangiocellular carcinoma of the liver in Japanese patients with hepatitis C virus-related cirrhosis. Cancer 2000; 88:2471– 7. 23. Han SL, Zhu GB, Yao JG, et al. Diagnosis and surgical treatment of primary hepatic cholangiocarcinoma. Hepatogastroenterology 2005; 52:348–51. 24. Shimoda M, Farmer DG, Colquhoun SD, et al. Liver transplantation for cholangiocellular carcinoma: analysis of singlecenter experience and review of the literature. Liver Transplantation 2001; 7:1023–33. 25. Valverde A, Bonhomme N, Farges O, et al. Resection of intrahepatic cholangiocarcinoma: a Western experience. J Hepatobiliary Pancreat Surg 1999; 6:122–7. 26. Kaczynski J, Hansson G, Wallersted S. Incidence, etiologic aspects and clinicophatologic features in intrahepatic cholangiocellular carcinoma: a study of 51 cases from a low-endemicity area. Acta Oncol 1998; 37:77–83. 27. Roayaie S, Guarrera J, Ye MQ, et al. Aggressive surgical treatment of intrahepatic cholangiocarcinoma: predictors of outcomes. J Am Coll Surg 1998; 187:365–72. 28. Sasaki A, Aramaki M, Kawano K, et al. Intrahepatic peripheral cholangiocarcinoma: mode of spread and choice of surgical treatment. Br J Surg 1998; 85:1206–9.
Ann. Surg. Oncol. Vol. 15, No. 7, 2008
29. Nakeeb A, Pitt HA, Sohn TA, et al. Cholangiocarcinoma. A spectrum of intrahepatic, perihilar, and distal tumors. Ann Surg 1996; 224:463–75. 30. Nozaki Y, Yamamoto M, Ikai I, et al. Reconsideration of the lymph node metastasis pattern (N factor) from intrahepatic cholangiocarcinoma using the international union against cancer TNM staging system for primary liver carcinoma. Cancer 2001; 83:1923–9. 31. Isa T, Kusano T, Shimoji H, et al. Predictive factors for longterm survival in patients with intrahepatic cholangiocarcinoma. Am J Surg 2001; 181:507–11. 32. Inoue M, Makuuchi T, Takayama , et al. Long-term survival and prognostic factors in the surgical treatment of massforming type cholangiocarcinoma. Surgery 2000; 127:498–505. 33. Grobmyer SR, Wang L, Gonen M, et al. Perihepatic lymph node assessment in patients undergoing partial hepatectomy for malignancy. Ann Surg 2006; 244:260–4. 34. Akatsu T, Shimazu M, Kawachi S, et al. Long-term survival of intrahepatic cholangiocarcinoma with hilar lymph node metastasis and portal vein involvement. Hepatogastroenterology 2005; 52:603–5. 35. Lang H, Sotiropulos GC, Fruhauf NR, et al. Extended hepatectomy for intrahepatic cholangiocellular carcinoma (ICC). Ann Surg 2005; 241:134–43. 36. Puhalla H, Schuell B, Pokorny H, et al. Treatment and outcome of intrahepatic cholangiocellular carcinoma. Am J Surg 2005; 189:173–7. 37. Slakey DP. Radiofequency ablation of recurrent cholangiocarcinoma. Am Surg 2002; 68:395–7. 38. Shirabe K, Shimada M, Tsujita E. Prognostic factors in nodenegative intrahepatic cholangiocarcinoma with special reference to angiogenesis. Am J Surg 2004; 187:538–42.
