This paper was presented at the 9th Congress of the European Hepato-Pancreato-Biliary Association, Cape Town, South Africa.
Short- and long-term results of extended left hepatectomy for colorectal metastases
Article first published online: 8 JUN 2011
© 2011 International Hepato-Pancreato-Biliary Association
Volume 13, Issue 8, pages 536–543, August 2011
How to Cite
Wicherts, D. A., de Haas, R. J., Andreani, P., Ariche, A., Salloum, C., Pascal, G., Castaing, D., Adam, R. and Azoulay, D. (2011), Short- and long-term results of extended left hepatectomy for colorectal metastases. HPB, 13: 536–543. doi: 10.1111/j.1477-2574.2011.00321.x
- Issue published online: 18 JUL 2011
- Article first published online: 8 JUN 2011
- Received 1 March 2011; accepted 25 March 2011
- colorectal metastases;
Background: An extended left hepatectomy is a complex hepatic resection often performed for large tumours in close relationship to major hilar structures. Operative outcomes of this resection for colorectal liver metastases (CLM) remain unclear. The aim of the present study was to assess short- and long-term outcome for patients with CLM after an extended left hepatectomy.
Methods: A retrospective analysis of consecutive patients undergoing an extended left hepatectomy for CLM in a large, single-centre cohort between January 1990 and January 2006 was performed.
Results: Thirty-one patients (3.9%) from a consecutive series of 802 patients who had undergone hepatic resection were identified as having met the definition of an extended left hepatectomy and were included for further analysis. Maximum tumour size was more than 60 mm in 15 patients, with a median size of 67.5 mm for the total group (range: 20 to 160 mm). Twenty-six patients presented with initially unresectable metastases, related to large tumour size in 11 patients and to a close relation with major vascular structures in six patients. Preoperative chemotherapy was administered to 29 patients. Combined vascular resection was performed in five patients. The mortality rate at 90 days was zero and post-operative morbidity occurred in 17 patients. R0 and R1 resections were performed in 17 and 11 patients, respectively. Three- and 5-year overall survival was 38% and 27%, respectively. Disease-free survival was 9% and 4% at 3 and 5 years. Morbidity did not differ between patients with and without a caudate lobectomy (9 of 17 patients vs. 8 of 14 patients, respectively) (P= 0.815).
Conclusions: An extended left hepatectomy for CLM can provide significant long-term survival. However, morbidity is increased in this complex procedure. A caudate lobectomy does not impact surgical outcome.
In current liver surgery, resections up to 70% of functional liver volume have been shown to be safe without the risk of post-operative liver failure.1 The most important indication for a major hepatectomy is the presence of an extensive intrahepatic tumour, either caused by multiple metastases or large tumour size. Extended left hepatectomies are, however, not comparable to any other liver resection because of their technical complexity related to a large resection surface, central liver dissection and resection of the caudate lobe. Nevertheless, patients with centrally located metastases with close contact to major vascular or biliary structures often require an extended left hepatectomy to achieve complete tumour clearance.
Recently, the authors have shown that major liver resections combined with a vascular resection and reconstruction can safely be performed in selected patients.2 However, extended left hepatectomies are still suggested to have higher morbidity and mortality rates compared with other resections because of the complexity of the procedure.3–5 Yet, the inclusion of mixed patient groups in all previous series interferes with an objective analysis of surgical morbidity after an extended left hepatectomy in patients with colorectal liver metastases (CLM).3,6–8 Different diseases with various underlying pre-existing conditions, such as cirrhosis in patients with a hepatocellular carcinoma or jaundice in patients with hilar tumours, can all influence surgical outcome.9 Critical analysis of associated risks and results of this challenging procedure is, however, crucial when offering it to patients as a potentially curative treatment.
The objective of the present study was to assess the short- and long-term outcome for patients with CLM after an extended left hepatectomy in a large, single-centre cohort. A secondary objective was to determine the impact of a caudate lobectomy on post-operative morbidity.
Patients and methods
All consecutive patients who underwent an extended left hepatectomy for CLM between January 1990 and January 2006 were included. Patients were selected from a prospectively maintained hepatobiliary database. Extended hepatectomies were defined as resections exceeding the boundaries of a standard left hepatectomy (segments II–IV).10,11 Extended left hepatectomies therefore included segments II–IV with segment I alone, or with segments V and VIII, or with segments I, V and VIII.
Systemic chemotherapy was administered to patients with technically unresectable liver metastases. Initial unresectability was defined as the inability to completely resect all liver metastases while leaving at least 30% to 40% of functional liver volume or the presence of extrahepatic disease. Local unresectability was defined either by involvement of major vascular structures such that a clear margin could not be achieved, involvement of multiple liver segments, or large tumour size such that the future remnant liver volume was deemed to be insufficient. During the study period, chemotherapy was also increasingly administered to patients with upfront resectable metastases to assess tumour chemo-responsiveness and to facilitate margin-negative resections.
Clinical response to chemotherapy according to WHO guidelines12 and more recently by using the response evaluation criteria in solid tumours (RECIST) criteria13 was assessed by computed tomography and/or magnetic resonance imaging in a multidisciplinary meeting with surgeons, oncologists and radiologists every four cycles of treatment. Surgery was considered only when complete resection of intra- and extrahepatic metastases could be achieved.
During surgery, the abdominal cavity was searched for extrahepatic metastases and suspicious lesions were examined by frozen section. Intra-operative ultrasound of the liver was used in all patients to confirm the number and size of metastases, to define their relation with intrahepatic vascular structures and to search for occult lesions. Liver resections were performed with the objective of radical resection, either achieved by single or multiple procedures. When tumour-free margins could not be obtained, owing to vascular proximity or the presence of multiple metastases, resection was still performed provided that it was macroscopically complete. Prophylactic abdominal drains were placed in all patients after a hepatectomy.
Portal vein embolization (PVE) and two-stage hepatectomy procedures were used accordingly for patients with estimated resections of more than 60% to 70% of liver parenchyma (based on pre-operative CT volumetrics) or with multiple metastases unable to be resected by a single hepatectomy.14,15 PVE was performed pre-operatively by a percutaneous or ileocolic technique decided by the treating surgeon.16,17 In selected patients, portal vein embolization and ligation was performed intra-operatively during resection of the primary tumour or the first step of a two-stage hepatectomy approach.15 Local ablative techniques (RFA or cryotherapy) were only used in combination to resection for three or less, otherwise unresectable metastases up to 3 cm in diameter in the future remnant liver. Vascular reconstruction was considered in cases with tumoural encasement of major vascular structures. For these resections, total vascular exclusion of the liver with or without venovenous bypass was generally used.2,18–20
R0 and R1 resections were defined by the absence (any tumour-free margin) or presence (tumour-free margin 0 mm) of microscopic tumour invasion of the resection margins, respectively.21
All post-operative complications occurring within 2 months after surgery were recorded and graded according to the Dindo–Clavien classification.22 An increase in the serum bilirubin level to >50 µmol/l and a decrease in prothrombin time to <50% on post-operative day 5 were used to define liver insufficiency.23 Post-operative mortality was assessed at 90 days.
Regular post-operative follow-up started at 1 month from hepatectomy, and then every 4 months, consisting of routine blood tests, serum tumour markers evaluation (CEA and CA19.9) and hepatic ultrasound. Thoracoabdominal CT imaging was performed every 8 months. The general policy was to continue chemotherapy after liver resection for six to eight cycles to reduce the risk of recurrence. Suspicion of tumour recurrence in the case of elevated tumour markers always necessitated radiological confirmation. Resection of intra- or extrahepatic disease recurrence was performed only when it could be globally curative and was generally preceded by chemotherapy to control the disease process.24
The χ2-test was used to compare categorical data (the number of patients with percentages) and continuous data (median with range) were compared using the Mann–Whitney U-test for non-parametric data. Survival curves were generated by the Kaplan–Meier method and compared using the log-rank test. Statistical significance was defined as P≤ 0.05. All statistical analyses were performed with SPSS version 13.0 (SPSS Inc., Chicago, IL, USA).
Between January 1990 and January 2006, a consecutive series of 802 patients underwent a partial hepatectomy for CLM with curative intent, 31 (3.9%) of which underwent an extended left hepatectomy (Fig. 1). The number of extended left hepatectomies performed between 1990–95, 1995–2000 and 2000–06 were 1, 10 and 20, respectively.
Demographics, primary and secondary tumour characteristics and chemotherapy regimens are provided in Table 1. Maximum tumour size was more than 60 mm at diagnosis in 15 patients. Intra- and post-operative variables are provided in Table 2. Fifteen patients received intra-operative red blood cell transfusions.
|Extended lefthepatectomy(n= 31)|
|Median age (range), years||59.5 (42–79)|
|Male/female||17 / 14|
|Colon/rectum||25 / 6|
|Median number (range)||3 (1–20)|
|Median size (range), mm||67.5 (20–160)|
|Cause of initial unresectability|
|Large tumour size||11|
|Close vascular relationship||6|
|Concomitant extrahepatic disease||5|
|Number of lines|
|Median number of cycles (range)||8 (3–27)|
|Last line regimen|
|Extended lefthepatectomy(n= 31)|
|Median number of detected metastases (range)||2 (1–14)|
|Anatomical and non-anatomical||18|
|Vascular resection and reconstruction||5|
|Inferior vena cava||1|
|Combined local treatment|
|Portal vein embolization||6|
|Median red blood cell transfusions (range), units||1 (0–28)|
|Mortality (≤90 days)||0|
|Median hospital stay (range), days||13 (8–31)|
In two patients with concomitant extrahepatic disease, lung metastases were scheduled to be resected after curative liver surgery. However, disease progression unable to be controlled by chemotherapy prevented lung surgery in both patients.
An R0, R1 and R2 margin occurred in 17, 11 and 0 patients, respectively. For three patients these data was missing. PVE did not affect the incidence of R1 resections (2 of 6 patients in the PVE group vs. 9 of 25 in the no PVE group) (P= 0.971).
Tumour recurrence was diagnosed in 26 patients after a median follow-up of 24.9 months (range: 6.0 to 85.2 months) for the total group. Median time to first recurrence was 9.8 months (range: 1.0 to 43.6 months). Intrahepatic recurrences occurred in 15 patients, which in seven patients were combined with extrahepatic recurrence. Three patients underwent a second liver resection. Localized extrahepatic disease recurrence was resected once in five patients, twice in one patient and three times in one patient.
At last follow-up, 20 patients have died, 4 patients are alive and 7 patients are lost to follow-up. Three- and 5-year overall survivals for the total group of 31 patients were 38% and 27%, respectively (Fig. 2). Overall survival did not differ according to R0-R1 status (36% and 54% at 3 years, respectively) (P= 0.165) or the presence of extrahepatic disease (26% and 41% at 3 years with and without extrahepatic disease, respectively) (P= 0.491). Disease-free survivals at 3 and 5 years were 9% and 4%, and were also not related to R0-R1 status (6% and 16% at 3 years, respectively) (P= 0.810).
Out of all 31 patients that underwent an extended left hepatectomy, 17 also had resection of the caudate lobe (segment I). Patient groups with and without a caudate lobectomy were similar in terms of age, gender distribution, primary and secondary tumour characteristics and chemotherapy treatment (Table 3).
|With a caudate lobectomy(n= 17)||Without a caudate lobectomy(n= 14)||P|
|Median age (range), years||61.7 (42–79)||56.8 (45–72)||0.874|
|Male/female||10 / 7||7 / 7||0.623|
|Colon/rectum||15 / 2||10 / 4||0.239|
|Median number (range)||2 (1–8)||4 (1–20)||0.287|
|Median size (range), mm||70.0 (20–160)||67.5 (20–100)||0.328|
|Cause of initial unresectability|
|Large tumour size||6||5|
|Close vascular relationship||5||1|
|Concomitant extrahepatic disease||4||1||0.217|
|Number of lines|
|Median number of cycles (range)||8 (4–18)||10 (3–27)||0.320|
|Last line regimen|
Of all five patients that underwent vascular resection and reconstruction in the total group, four had resection of the caudate lobe (P= 0.217). Total vascular exclusion was used in six patients with a caudate lobectomy and in three patients without a caudate lobectomy (P= 0.338). Intra-operative red blood cell transfusions were necessary in 10 patients that underwent a caudate lobectomy (compared with five patients in the control group) (P= 0.198).
Post-operative morbidity occurred in nine and eight patients with and without a caudate lobectomy, respectively (P= 0.815). Although R1 resections were more frequent in the caudate lobectomy group because of vascular proximity of the metastases (8 vs 3 patients) (P= 0.102), recurrences occurred in a similar proportion of patients in both groups (in 14 and 12 patients with and without caudate lobectomy, respectively) (P= 0.886).
Owing to significant development of surgical techniques as well as extended indications for a partial hepatectomy in patients with CLM, major hepatectomies are now increasingly being performed. Extended left hepatectomies, however, represent a specific category of liver surgery owing to the technical difficulty related to the dissection of hilar structures as well as resection of the caudate lobe. Until now, specific analysis of short-term outcome of this procedure in a sufficient cohort of patients with CLM was lacking.3,4,7,8
The main result of the present study is that extended left hepatectomies for patients with CLM can result in significant long-term survival. Three-year survival could be achieved in 38% of patients, not so different from general survival rates reported in the literature.25–27 Post-operative mortality was nil, with however a significant morbidity rate. Of note, disease-free survival was only 9% at 3 years and 4% at 5 years. A possible explanation is the frequency of R1 resections associated with this type of resection (11 of 31 patients) and a high incidence of disease recurrence. For metastases in close contact to vascular structures, tumour-free resection margins are often impossible to achieve. Nevertheless, overall survival was not influenced by a R1 resection. As previously reported, in the current era of effective chemotherapy with the possibility of a repeat hepatectomy, R1 resections do not affect survival.28 This furthermore confirms the value of extended left hepatectomy for patients with complex metastases otherwise unresectable. Complete macroscopic removal of all lesions remains, however, essential.14,15
Previous reports describing extended left hepatic resections all included mixed populations of patients with cholangiocarcinomas (including patients with jaundice), hepatocellular carcinomas (usually in cirrhotic patients) and colorectal metastases.3,4,7,8 Owing to the inclusion of small patient numbers, information on morbidity and mortality rates for patients with CLM is absent or very limited. The present study, to the authors' knowledge, is the first to specifically address short-term outcome of an extended left hepatectomy in a homogeneous group of patients with CLM only.
Patients usually had central, large metastases in close relation with major hilar structures, reflecting the technical complexity of this procedure. Resection of centrally located metastases in close proximity with main vascular structures often needs vascular resection and reconstruction with total vascular exclusion to control intra-operative blood loss. Therefore, operative characteristics of extended left liver resections generally differ from those of major right-sided resections. In a natural way, colorectal liver metastases are most frequently located in the right lobe, owing to anatomical differences in weight and portal venous inflow of both hemilivers.29,30 This natural distribution of metastases results in a tumour presentation which is more often multifocal in patients selected for a right hepatectomy. Right-sided resections usually include more liver segments, and additional strategies such as PVE and two-stage approaches are more frequently required to achieve complete tumour resection. Indeed, the authors' recent report describing the results of PVE also mainly included right-sided resections.14 On the contrary, vascular resection is much less frequently performed in extended right hepatectomies (8% of patients in the authors' own series; data not shown).
Biliary leakage and (non)infected collections occurred frequently in this series. Previous reports also found that bile leakage was a frequent complication in left hepatectomy patients because of the difficult parenchymal transection.3,4,7 The high incidence of collections could be related to the difficulty to have an efficient drainage after a left-sided hepatectomy compared with right-sided resections. Patients undergoing right hepatectomies have a higher incidence of post-operative liver insufficiency because of the resection of higher liver volumes.14,15 In combination with massive pre-operative chemotherapy, the risk of liver failure is known to be elevated.31,32 Post-operative liver insufficiency, on the other hand, occurred only in two patients in the present study.
Liver metastases involving the caudate lobe add to the complexity of a curative hepatectomy caused by its difficult anatomical localization. The caudate lobe lies in close relation with the inferior vena cava, the portal bifurcation and the confluence of the left and middle hepatic veins.33 Its mobilization from the inferior vena cava with the ligation of its short hepatic veins during resection increases the risk of intra-operative blood loss and associated morbidity. However, owing to surgical refinements, resection of the caudate lobe is increasingly performed.34 The fact that surgical margin width is of limited influence on survival has also contributed to this development.28,35
In the present study, patients that underwent a caudate lobectomy often presented with initially unresectable metastases as a result of vascular proximity. Combined vascular resection was frequently performed, but surgical morbidity was not increased compared with a control group without a caudate lobectomy. Recurrences occurred in a similar proportion as in patients without caudate lobe involvement, although patients with a caudate lobe resection more often presented with R1 resections. Khan et al. found similar results, confirming that patients with metastases in the caudate lobe can be offered a similar short-term outcome as other patients with CLM.34 Others suggested a decreased survival in the case of caudate lobe involvement as a result of a early spreading of tumour cells through direct venous drainage in the vena cava.36,37 This was, however, not confirmed in the present cohort.
In conclusion, an extended left hepatectomy for CLM is technically challenging with a high risk of microscopically involved margins and tumour recurrence. However, in a multidisciplinary setting, it may provide significant long-term survival. Owing to the complexity of this procedure with a high morbidity rate and the so-called volume–outcome relationship in this field, patients should be referred to specialized liver surgery centres.
Conflicts of interest
- 10Terminology committee of the IHPBA. (2000) The Brisbane 2000 terminology of liver anatomy and resections. HPB 2:333–339.
- 111957) Le foie: études anatomiques et chirurgicales. Paris: Masson et Cie.. (
- 12World Health Organization. (1979) Handbook for Reporting Results of Cancer Treatment. Geneva: World Health Organization.
- 132000) New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst 92:205–216., , , , , et al. (
- 212002) TNM Classification of Malignant Tumours. Hoboken, NJ: John Wiley & Sons., . (