Local surgical resection of hilar cholangiocarcinoma: Is there still a place?

Abstract

In recent decades, surgical treatment of hilar cholangiocarcinoma has moved toward liver surgery in association with biliary resection in order to increase radicality and to achieve better survival. Results of local resection compared with hepatectomy associated with bile duct resection and its actual indications have to be clarified. A systematic review of relevant studies published before December 2007 was performed. Original published studies comparing the results of isolated local excision with those of hepatectomy associated with bile duct resection were identified and the reported results were synthesized. The pathologic data suggest that isolated bile duct resection cannot be adequate: required wide surgical margins; neoplastic extension along perineural sheaths; Segment 1 neoplastic invasion. Considering postoperative outcomes, in the 1990s, local resection had significantly lower mortality rates than liver resection. In recent years, the short-term results of liver surgery have improved significantly, while mortality rates have decreased. The R0 resection rate is significantly higher after associated liver resection. Comparison of survival results between local resection and associated liver surgery is difficult because, in the majority of series, the treatment was planned according to tumor extension. Better long-term outcomes have been reported after liver resection than after isolated bile duct resection, even for Bismuth-Corlette type I–II cholangiocarcinoma. Long-term survivors after local resection have been reported in a few selected patients with Bismuth–Corlette type I Tis-T1 or papillary neoplasm.

Introduction

The surgical treatment of hilar cholangiocarcinoma has changed completely in recent decades: before 1980, the majority of patients were not resected, and in a few cases local excision of the tumor was performed with low radicality and poor long-term outcome. Since 1980, indications for resection have progressively improved and liver resection has been associated with bile duct resection in order to increase radicality and achieve better survival results ¹,²,³,⁴. Now, local resection seems to have a very narrow role, but its results compared with those of hepatectomy associated with bile duct resection and its actual indications have to be clarified. We have tried to answer these questions in this systematic review of the literature.

Methods

Original published studies were identified by searching the MEDLINE database (up to December 2007). Articles were selected using the key words ‘cholangiocarcinoma’, ‘local resection’, ‘hepatic resection’, ‘surgery’ to identify all reports that could pertain to surgical management of hilar cholangiocarcinoma. Manual cross-referencing was performed and relevant references from selected articles were reviewed.

Results

No randomized trials are available. All the data were collected from retrospective studies and from a few recent prospective studies, including a small number of patients. Many pathologic data suggest that isolated bile duct resection is inadequate in the treatment of Klatskin tumor. Cholangiocarcinoma is an aggressive neoplasm: extension along the perineural sheath and associated lymphangiosis carcinomatosa are common ⁵,⁶. The definition of “early bile duct cancer” is not widely accepted because in the bile duct the submucosal layer is thin and often lacking, and the muscularis mucosae is absent; even small tumors may behave in an aggressive and invasive pattern ⁷. In 1999, Kayahara reported neural invasion in no patients with mucosal tumor (Tis) and in 33% of patients with T1 tumor ⁶. In 2002, Ebata reported that cholangiocarcinoma had both superficial and intramural extension; to remove invasive neoplasm a 10 mm margin was necessary and 20 mm in the case of all non-invasive components ⁸. These wide margins can be achieved only by liver resection when cholangiocarcinoma involves right or left hepatic ducts. Moreover, Sg1 is often invaded by the tumor; a high rate of tumor recurrence into Sg1 has been reported after isolated bile duct resection ²,⁹,¹⁰,¹¹,¹²,¹³. Sg1 may be involved by its bile ducts invasion, direct infiltration or perineural extension ²,⁹,¹⁰,¹¹,¹²,¹³. In order to compare isolated local resection with bile duct excision associated with liver resection, three topics have to be discussed: short-term outcomes, radicality, and survival (Table I).

Table I. Comparison between local resection (LR) and bile duct resection with associated hepatic resection (HR) in published series.

		#		Mortality (%)			Radicality (%)			5-year survival (%)			Median survival (months)
Author	Year	LR	HR	LR	HR	p	LR	HR	p	LR	HR	p	LR	HR	p
Boerma 17	1990	201	188	80.0	15.0	<0.05				70.0	17.0	< 0.05
Bismuth 35	1992	10	13	0	0	n.s.	30.0	46.1	n.s.
Baer 18	1992	12	11	0	80.3	n.s.							36	32	n.s.
Pichlmayr 15	1996	30	95	30.3	12.7	n.s.				28.9	26.3	n.s.
Miyazaki 16	1998	11	65	0	15.0	n.s.	45.0	75.0	<0.05	16.0 3 yrs	33.0 3 yrs	n.s.
Launois 34	1999	11	25	0	16.0	n.s.				27.3	60.0	< 0.05
Neuhaus 5	1999	14	66	0	90.1	n.s.	29.0	61.0	<0.05	0	28–57*	n.s.
Kosuge 32 **	1999	13	52	70.7	90.6	n.s.	38.5	55.8	n.s.				38.6	27.0	n.s.
Nimura 39	2000	8	100	0	60.0	n.s.				16.0	26.0	n.s.
Launois 19(French survey)	2000	51	47	14.0	17.0	n.s.							23	24	n.s.
Jarnagin 30	2001	18	62	60.0	11.0	n.s.	56.0	84.0	<0.05	0	37.0	< 0.05
Capussotti 25	2002	4	32	0	30.0	n.s.	75.0	90.6	n.s.	0§	54.5§	< 0.05§
Kondo 23	2004	9	31	0	0	n.s.	100	100	n.s.				20.8§§	NA§§	<0.05§§
Jang 31	2005	25	23	0	0	n.s.				28.0	47.8	n.s.
Dinant 3	2006	60	37	13.1	26–17‡	n.s.	14.8	42–37‡	<0.05				21.5	38–69‡	n.s.

*Different survival rates depending on type of hepatectomy.

**Local resection/limited hepatic resections vs major hepatectomies.

§Only BC I–II included.

§§Local resection/caudate segmentectomy/left hepatectomy vs right hepatectomy.

‡Results after right hepatectomy–left hepatectomy.

Short-term outcomes

Local resection has been widely considered as safer than liver resection and with lower mortality rates ¹⁴,¹⁵,¹⁶. In 1990, Boerma reviewed published papers and collected 581 patients undergoing resection for Klatskin tumor: considering the most relevant series, mortality was significantly lower after bile duct resection than after hepatectomy (8% vs 15%) ¹⁷. These data were confirmed in the 1990s: in 1992, Blumgart's group reported no mortality and 25% morbidity after local excision compared with 8% and 36%, respectively, after extended procedures ¹⁸; in 1996, Pichlmayr reported mortality rates of 12.7% after liver surgery associated with bile duct resection vs 3.3% after local excision ¹⁵. The increased mortality rate after liver resection was considered a clear indication of local resection: even though associated liver surgery could improve radicality, long-term benefits were lost because of high mortality rates ¹⁴.

In 2000, Launois published different results: in this French survey, mortality rates were high but similar in patients with and without liver resection (17% vs 14%) ¹⁹. Postoperative outcomes of liver surgery have improved significantly in recent years, thanks to better patient selection, preoperative biliary drainage, and portal vein embolization ²⁰,²¹,²²,²³. In 2000, Tsao compared the results of surgical treatment in a Japanese center (Nagoya), where liver resection was performed routinely, with those of an American center (Lahey clinic), where isolated bile duct resection was preferred: short-term outcomes were good and similar between the two groups, i.e. 4% vs 8% mortality and 44% vs 51% morbidity ²⁴. In 2002, our group published an article on 36 patients resected for hilar cholangiocarcinoma, and with associated liver resection in 32 (88.9%): mortality and morbidity rates were 2.8% and 47.2%, respectively ²⁵. Some recent articles by oriental groups have reported no mortality after bile duct resection associated with hepatectomy ²⁰,²²,²³.

Radicality

According to the pathologic characteristics of Klatskin tumor, local excision is inadequate for radical resection. This has been confirmed in surgical series: the rate of radical resections has increased along with the rate of associated liver resections ¹. R0 resection rates ranged from 15% in Cameron's series with 20% hepatectomy to 56% in Blumgart's series with 60% liver resections, reaching 80% when liver resection was associated in about 80% of cases ²⁶,²⁷,²⁸,²⁹. Nimura performed liver resection in 98% of cases, with a radicality rate of 83% ¹¹. Our center had similar results: liver resections in 89% of patients and a radicality rate of 89% ²⁵. Many studies have reported that R0 resections are significantly less common after local excision than after bile duct resection associated with liver surgery ³,⁵,¹⁶,³⁰. In 2000, Tsao compared between the results of Klatskin tumor resection at Nagoya University and the Lahey Clinic: the first center performed liver resection in 89% of cases and R0 resection was achieved in 79%; the second performed hepatectomy in only 16% of patients and radicality decreased significantly to 28% ²⁴. Liver resection is not always necessary to reach negative margins: in 2004, Kondo published a series of 40 consecutive patients with radical resection, including 9 cases treated by isolated bile duct resection ²³.

Survival

In the majority of published series, no evidence has been found of any statistical difference in survival between local resection and extended surgery ²,⁴,¹⁴,¹⁵,¹⁶,¹⁸,¹⁹,³¹,³²,³³. In 1999, Launois published a series of 40 consecutive resected patients: survival after local resection was significantly better than after associated hepatectomy (5-year survival rates 27% vs 6%) ³⁴. These data could be related to the fact that the treatment was planned according to tumor location and that liver resection was scheduled for patients with more extended disease. In 2000, Tsao, on comparing oriental and US experiences, reported significantly better survival in Japanese patients undergoing more aggressive surgical strategy (5- and 10-year survival rates were 16% and 12% vs 7% and 0%) ²⁴.

Some studies have reported significantly increased survival after associated liver resection: 1) In the Boerma review published in 1990, survival was significantly lower after local resection than after extended surgery (5-year survival rate 7% vs 17%) ¹⁷. 2) In 2001, Jarnagin analyzed 80 consecutive patients. Five-year survival rates were 37% in patients with liver resection and 0% in patients without it. In order to exclude the possibility that the difference in survival was related to radicality rate, the analysis was repeated including only R0 patients and the result was confirmed ³⁰. 3) In 2004, Kondo reported long-term results in 40 consecutive patients with R0 resection. Nine patients treated by isolated bile duct resection had significantly decreased survival compared with 17 patients undergoing right hepatectomy ²³.

All these data concern hilar cholangiocarcinoma with different extension into the bile ducts. Patients undergoing local resection probably had tumors with or without minimal involvement of the bile duct confluence, but this is not clearly defined in the majority of articles. In order to clarify whether local resection may have a role in these patients, we focused our analysis on reported results in Bismuth–Corlette (BC) type I–II hilar cholangiocarcinoma.

Two articles from France suggest that local resection could be indicated in selected patients ³⁴,³⁵. In 1992, Bismuth reported 3 patients with BC I tumor undergoing isolated bile duct resection: 2 cases with R0 resection had long-term survival without recurrence; 1 case with R1 resection had long-term survival with recurrence and re-resection ³⁵. In contrast, all the patients with local resection for BC II tumor had recurrence. In 1999, Launois reported 4 patients with BC I (undergoing local resection in 3 cases) and 4 patients with BC II neoplasms (local resection in 3) ³⁴. Five-year survival results were good (type I 20%; type II 25%), similar to those reported for more extended BC types. Patients with local resection mainly had Tis and T1 tumors. Two long-term survivors have been reported: one BC I T1bN0M0 and one BC II TisN0M0.

Different results have recently been published with poor outcomes. In the Neuhaus series published in 1999, radicality of isolated local resection was 33% (2/6) in BC I and 25% (1/4) in BC II tumors; among them, no patient survived 5 years ⁵. Similar results have been reported by Su after local resection in 1996 (25% R0 resection rate for BC I) and by van Gulik in 1999 (19% R0 resection rate for BC I–II) ³⁶,³⁷. In our experience published in 2002, R0 resection was achieved in 2 out of 3 BC I–II tumors treated by local resection, but patient survival was significantly lower after isolated bile duct resection (3 cases) than after associated liver resection (6 cases), and no patient with bile duct resection was alive at 2 years ²⁵.

In 2005, Jang reported 25 patients undergoing isolated bile duct resection for BC I–II or common hepatic duct cancer ³¹. Seven patients (28%) survived 5 years or more, but three of them were alive with recurrence despite their early stage (T1N0 in 2 and T2N0 in 1).

In Kondo's series, 9 patients (6 BC I and 3 BC II) underwent local resection: even though all had radical resection, their survival rate was significantly lower than that of patients with associated liver surgery (median survival 19 vs 21 months) ²³. In 2003, Makuuchi's group reported results of systematic liver resection: mean survival was 42 months in 9 patients BC I and 51 months in 8 BC II ²².

A recent article by the Nagoya group focused on 54 patients affected by BC type I–II tumors ³⁸. In 14 cases, an isolated bile duct resection was performed. Unfortunately, the study analyzed the outcomes of right hepatectomies with caudate lobectomy compared with those of more limited resections, and few data are available on isolated bile duct resection without hepatectomy. Accordingly, the authors suggested a surgical approach based on cholangiographic tumor type: extended hepatectomy is always necessary in the event of nodular or infiltrating tumor; on the contrary, bile duct resection with or without limited hepatectomy can be adequate in the case of papillary tumor without superficial cancer spreading. Regardless, further studies are needed to better define these indications, mainly because only two patients of the papillary group have been treated by isolated local resection.

Consensus statements

Local resection is not an adequate treatment for hilar cholangiocarcinoma involving the bile duct confluence; associated liver resection should be recommended. In Bismuth–Corlette type I hilar cholangiocarcinoma, benefits of survival by association of biliary and liver resection have been reported, but further studies are needed. Local resection should be scheduled only for small papillary Klatskin tumors without bile duct confluence involvement (type I) confined to the bile duct wall (Tis and T1). Extension of treatment should always be determined in accordance with the patient's condition.