Current Colorectal Cancer Reports

, Volume 9, Issue 2, pp 203–212

Management of the Hepatic Lymph Nodes During Resection of Liver Metastases from Colorectal Cancer: A Systematic Review

Authors

  • Rahul S. Koti
    • University Department of SurgeryRoyal Free Campus, UCL Medical School
  • Constantinos Simillis
    • University Department of SurgeryRoyal Free Campus, UCL Medical School
  • Kurinchi S. Gurusamy
    • University Department of SurgeryRoyal Free Campus, UCL Medical School
  • Michael Jacovides
    • University Department of SurgeryRoyal Free Campus, UCL Medical School
    • University Department of SurgeryRoyal Free Campus, UCL Medical School
Colorectal Cancer Hepatic Metastases (KK Tanabe, Section Editor)

DOI: 10.1007/s11888-013-0165-6

Cite this article as:
Koti, R.S., Simillis, C., Gurusamy, K.S. et al. Curr Colorectal Cancer Rep (2013) 9: 203. doi:10.1007/s11888-013-0165-6
  • 110 Views

Abstract

The aim of this review is to assess the role of hepatic lymphadenectomy in patients undergoing liver resection for colorectal metastases. Meta-analysis of trials identified by a systematic literature search of the Medline, Embase and Central databases was performed. There were no randomized controlled trials which assessed the survival benefit to patients undergoing liver resection for colorectal metastases of either a routine or a ‘selective’ lymphadenectomy. The prevalence of nodal metastases after lymphadenectomy was 8.4 %. The overall 3-year and 5-year survival rates in patients with hepatic lymph node metastases undergoing hepatectomy were 21.8 % (63 of 288 patients) and 8.5 % (27 of 315 patients), respectively, compared with 58.2 % (1,366 of 2,346 patients) and 47.5 % (1,717 of 3,609 patients) in patients undergoing hepatectomy who had no hepatic nodal metastases. The odds ratios for 3-year and 5-year mortality in node positive disease compared with node negative disease were 4.54 (95 % confidence interval 3.15–6.54) and 6.33 (95 % confidence interval 4.28–9.36), respectively. In conclusion, long-term survival rates are low in patients undergoing hepatectomy with hepatic lymph node metastases. The poor outcome is irrespective of whether the nodal metastases are discovered following routine lymphadenectomy or are detected because of pathological enlargement. Further trials in this patient group are required.

Keywords

LiverColorectal cancerMetastasesResectionHepaticLymph nodesHilarPortalPara-aorticCoeliacSurvivalSurgeryChemotherapySystematic reviewMeta-analysis

Introduction

Liver resection for colorectal cancer liver metastases in the presence of hepatic lymph node involvement is controversial. Metastatic involvement of hepatic lymph nodes is considered one of the worst prognostic factors [1, 2] and has been labelled as a contraindication to surgery [2]. Metastatic involvement of hepatic lymph nodes has been shown to be an independent predictor of poor outcome, with 5-year survival rates after liver resection, despite resection of involved lymph nodes, ranging between 0 and 4.3 % [35]. The alternative to surgery for patients with resectable liver metastases in the face of known hepatic nodal metastases is to undergo systemic chemotherapy. In the past, with 5-fluorouracil and leucovorin based chemotherapy for colorectal cancer liver metastases, the response rate was 20–30 %, with median survival time of 11–12 months [6•]. With the advent of more effective systemic agents such as oxaliplatin, irinotecan, cetuximab and bevacizumab, improved response rates of 33–62 %, median overall survival time of more than 20 months and improved progression-free survival are now seen in patients with liver metastases from colorectal cancer [6•]. These developments in modern chemotherapy may have an impact on results of surgical resection in patients with extrahepatic disease, and with more effective therapies there may be an indication for both liver resection and lymphadenectomy in those with chemotherapy-sensitive disease.

Adam et al. [7•] reported a 5-year disease-free survival rate of 11 % and a 5-year overall survival rate of 18 % after combined liver resection and lymphadenectomy in their prospective cohort of patients who had preoperatively diagnosed hepatic lymph node involvement and had responded to neoadjuvant chemotherapy. They also described a difference in survival dependent on the site of nodal metastases. The overall 5-year survival rate for patients who had metastases in the hepatic pedicle lymph nodes and underwent hepatectomy and lymphadenectomy was better (25 %) than for those undergoing hepatectomy along with resection of involved coeliac and para-aortic lymph nodes (0 %). Jaeck et al. [3] reported that in their cohort of 17 patients in whom lymph node involvement was diagnosed after routine lymphadenectomy during liver resection, the 3-year survival rate was 38 % in patients who had nodal involvement in the hepatoduodenal ligament compared with 0 % in patients who had nodal involvement around the hepatic artery and coeliac trunk. However, the study did not demonstrate whether the outcome was better overall in those undergoing lymphadenectomy, and no survival advantage was shown with routine lymphadenectomy. Pulitanò et al. [8•], in a cohort of 61 patients who underwent liver resection for colorectal liver metastases, most of whom had received preoperative as well as adjuvant chemotherapy, found an overall 5-year survival rate of 18 % in patients with lymph node involvement, with the survival difference associated with the location of the nodal metastases. These studies have shown that long-term survival can be achieved in highly selected patients undergoing liver resection along with resection of hepatic lymph node metastases. However, the number of patients in these studies is small and the opportunity is high for falsely positive conclusions.

In a relatively large series of 127 patients undergoing liver resection for colorectal liver metastases with concomitant resection of extrahepatic disease, patients with hepatic portal lymph node metastases had worse survival than those with lung or ovarian metastases from colorectal cancer [9•]. Intraoperative detection of hepatic lymph node metastases is another factor which has been shown to have a worse outcome [7•, 9•], and survival rates were lower than those if it was diagnosed preoperatively [7•]. This seems counterintuitive since one would expect preoperative imaging to detect more extensive disease than intraoperative examination, and hence intraoperative detection would be expected to have a relatively better prognosis. What is behind this difference needs to be explored further. Also, does intraoperative detection of hepatic lymph nodes have a worse outcome than not detecting them? This again has not been explored.

When macroscopic nodal involvement (enlarged, firm or obvious tumour infiltration) is not evident preoperatively or intraoperatively, nodal involvement (micrometastases) may be detected on routine node sampling. Microscopic hepatic lymph node involvement (found on histology of the resected lymph nodes and liver specimen) has been shown to have a negative impact on survival [1, 3, 5, 10]. In the modern era of more effective chemotherapy regimens, routine hepatic lymphadenectomy concurrent with liver resection for colorectal metastases could have a role in improving staging to guide adjuvant chemotherapy regimens [10, 11]. However there is no consensus on whether routine hepatic lymphadenectomy or node sampling should be performed [10, 11].

The aims of this review are to determine in patients with resectable colorectal liver metastases:
  • Whether there is a survival benefit from routine hepatic lymphadenectomy

  • Whether there is a survival benefit from lymphadenectomy in patients with proven hepatic nodal metastases

  • Whether the extent of lymphadenectomy influences survival

  • The reported risks of lymphadenectomy under these circumstances

We published a systematic review of the literature on hepatic lymph node involvement in colorectal metastases in 2007 [12]. We have conducted a further literature analysis to update the review.

Methods

Search Strategy

The following databases were searched: Medline (January 1990 to November 2012), Embase (January 1990 to November 2012) and Central (Cochrane Register of Controlled Trials in the Cochrane Library; issue 1, 2012 from 1990). The following search strategy was used: ("Neoplasm Metastasis"[MeSH] OR metasta* OR secondar* OR spread OR cancer OR carcinoma OR tumour Or tumor OR neoplasm) AND (colon Or colonic OR colorect* OR rectal OR rectum OR gut OR intestine OR bowel OR "Intestine, Large"[MeSH] OR "Colorectal Surgery"[MeSH] OR "Intestinal Neoplasms"[MeSH]) AND (("Liver"[MeSH] OR "Liver Neoplasms"[MeSH] OR "Liver Diseases"[MeSH] OR liver OR hepatic) AND (segmentectomy OR resection) OR "Hepatectomy"[MeSH]) AND "humans"[MeSH Terms] AND English[Lang] AND ("1990"[PDAT] : "3000"[PDAT]).

Equivalent search strategies were used in Embase and Central.

Inclusion and Exclusion Criteria

The following studies were included: randomized controlled trials; comparative series; published in journals from January 1990 onwards; full text in English language; should have survival or cancer recurrence data; and a minimum duration of follow-up of 1 year.

The following studies were excluded: number of patients undergoing liver resections below 10; includes liver resection for other cancers (primary or secondary); includes repeat or multistaged liver resections; includes combined excision of liver and lung metastases; hepatic lymph node status not reported; mention of extrahepatic disease but not clearly stated as involving hepatic lymph nodes; mention of regional lymph nodes but not clear whether the lymph nodes drain the primary tumour or the secondary tumour; not possible to identify the survival or recurrence data for hilar node positive and negative status separately; no controls (e.g. survival reported only in node negative disease and node positive patients did not undergo resection); lost to follow-up more than 10 %.

The following outcomes were measured: in-hospital mortality/surgical mortality/30-day mortality; 1-, 3- and 5-year mortality.

Data Extracted

The following data were extracted from each study using a custom-designed data extraction form: year of publication; year of study; country of study; type of study; population characteristics such as age and gender ratio; inclusion and exclusion criteria used in individual studies; number of individuals entering the study; follow-up period; lost to follow-up; tests performed during follow-up period; hepatic lymph node involvement and how it was diagnosed; any adjuvant therapy (including neoadjuvant therapy)—indication and details; whether routine hepatic lymphadenectomy was performed; operating time; blood loss, number of units transfused; hospital stay; complications of surgery (directly related to surgery—such as bleeding, bile leak, intra-abdominal collections, etc. and general complications such as pneumonia; cardiac complications, etc.); survival and mortality data; time to recurrence; type of recurrence (including local or regional or remote in relation to liver and the actual sites of recurrence); and quality of life measures (however described by the author).

Statistical Methods

Review Manager 5 [13] was used for meta-analysis. The odds ratio (OR) with 95 % confidence interval (CI) was calculated for the binary outcomes 1-year, 3-year, and 5-year mortality. The hazard ratio (HR) with 95 % CI was calculated for the time-to-event outcome ‘overall mortality’. If the HR with 95 % CI was not reported in the publications, the values were calculated from other information and the Kaplan–Meier curve using methods described by Parmar et al. [14]. The random-effects model [15] and the fixed-effect model [16] were used. The Mantel–Haenszel method was used to combine the binary outcomes and the generic inverse variance method was used to combine the time-to-event outcomes. Subgroup analysis, based on whether the studies included nodes involved macroscopically or microscopically, was performed. Other subgroup analyses performed were those based on whether routine lymphadenectomy or selective lymphadenectomy was performed; whether chemotherapy was used, and whether only the lymph nodes along the hepatoduodenal ligament were removed (as compared with the removal of the other groups of nodes).

Definitions Used

The following definitions were used in this review:
  • Hepatic lymph nodes. The term ‘hepatic’ has been used to maintain consistency and includes all subgroups of hepatic regional lymph nodes. The different terms used by different authors to define subgroups were as follows: hilar nodes, portal nodes, hepatoduodenal ligament nodes, hepatic pedicle nodes (nodes along the hepatoduodenal ligament, retropancreatic, common hepatic artery and celiac artery), perihepatic nodes, retropancreatic nodes, pedicular nodes (distal to gastroduodenal artery branch), common hepatic artery nodes, celiac nodes, para-aortic nodes and regional nodes.

  • Primary site. Primary site of origin in the colon or rectum

  • Local recurrence. Recurrence at the site of liver resection. This has no relation to the recurrence of the tumour at the primary site.

  • Regional recurrence. Perihepatic area including porta hepatis. Again, this has no relation to the regional lymph nodes draining the primary tumour site.

  • Remote recurrence. Recurrence of cancer in sites not included in the above two categories. In most cases (except in hepatic flexure tumours), recurrence at the primary site will be included in this category.

  • Macroscopic lymph node involvement. Lymph node involvement as detected radiologically (preoperatively or peroperatively) or by visual and tactile assessment.

  • Microscopic lymph node involvement. Lymph node involvement not detected radiologically (preoperatively or peroperatively) or by visual and tactile assessment but detected by resection followed by microscopic examination.

  • Lymph node involvement. Macroscopic or microscopic lymph node involvement or both.

  • No lymph node involvement. Neither macroscopic nor microscopic lymph node involvement (if assessed).

  • Routine lymphadenectomy. Lymphadenectomy performed routinely in the presence or absence of lymph node involvement.

Results

A total of 3,837 references were identified after exclusion of duplicate references through electronic searches of Pubmed, Embase and Central. Of these, 3,795 references were excluded after review of the title and abstract. The remaining 42 studies were reviewed in detail, and 21 of them were excluded because they did not meet the inclusion criteria. There were no randomized controlled trials comparing liver resection alone with liver resection along with routine lymphadenectomy or lymphadenectomy in patients with involved nodes. Similarly, there were no randomized controlled trials comparing liver resection with non-surgical treatments such as chemotherapy in patients with node positive disease who were otherwise suitable for liver resection.

Thus, 21 studies [35, 7•, 8•, 9•, 11, 1730] involving 4,618 patients undergoing liver resection for colorectal metastases were included for meta-analysis: 391 patients were node positive and 4,227 were node negative (Table 1).
Table 1

Characteristics of included studies

Authors and year of publication

Total number of patients

Number (percentage) of node positive cases

Routine or selective lymphadenectomy

Microscopic or macroscopic node involvement or both

Group of nodes dissected

Chemotherapy

Indications for chemo therapy

Ambiru et al. [17] 1999

149

8 (5.4 %)

Routine

Both

Hepatoduodenal ligament

Adjuvant local or none

Not stated

Minagawa et al. [24] 2000

235

6 (2.6 %)

Selective

Both

Hepatic hilum

Not stated

Not stated

Carpizo et al. [9•] 2009

127

27 (21.3 %)

Not stated

Both

Portal lymph nodes (hepatoduodenal ligament and hepatic artery)

Adjuvant local or systemic or none

Not stated

Yasui et al. [22] 1995

52

8 (15.4 %)

Routine

Both

Hepatic pedicle (hepatoduodenal ligament, retropancreatic, hepatic artery)

Not stated

Not stated

Jonas et al. [27] 2007

204

27 (13.2 %)

Routine

Both

Hepatoduodenal ligament

Adjuvant or neoadjuvant or none

Not stated

Settmacher et al. [28] 2011

382

18 (4.7 %)

Not stated

Not stated

Not stated

Adjuvant or neoadjuvant or none

Not stated

Lupinacci et al. [30] 2013

26

5 (19.2 %)

Routine

Microscopic

Hepatic hilum

Adjuvant

All

Oussoultzoglou et al. [26] 2008

132

12 (9.1 %)

Routine

Both

Hepatic pedicle

Adjuvant or neoadjuvant or none

Not stated

Adam et al. [7•] 2008

763

47 (6.2 %)

Selective

Both

Pedicular, coeliac, para-aortic

Selective neoadjuvant and adjuvant

Poor prognostic factors

Beckurts et al. [4] 1997

126

35 (27.8 %)

Routine

Both

Hepatoduodenal ligament

Not stated

Not stated

Pulitanò et al. [8•] 2012

1,519

61 (4.0 %)

Selective

Both

Hepatoduodenal ligament and retropancreatic area, common hepatic artery and coeliac axis, para-aortic

Neoadjuvant and adjuvant

Not stated

Harms et al. [21] 1999

155

39 (25.2 %)

Routine

Both

Hepatoduodenal ligament

Adjuvant local or systemic or none

Poor prognostic factors

Laurent et al. [5] 2004

156

23 (14.7 %)

Routine

Microscopic

Hepatic pedicle (hepatoduodenal, hepatic artery)

Adjuvant systemic

Not stated

Bennett et al. [11] 2008

45

8 (13.6 %)

Routine

Microscopic

Portocaval, pancreaticoduodenal, common hepatic artery

Selective neoadjuvant or adjuvant

Not stated

Kokudo et al. [18] 1998

94

7 (7.4 %)

Not stated

Not stated

Hepatoduodenal ligament

Adjuvant local or systemic or none

Not stated

Sanchez-Cespedes et al. [23] 1999

16

8 (50 %)

Not stated

Microscopic

Perihepatic lymph nodes

Not stated

Not stated

Nakamura et al. [19] 1992

22

6 (27.3 %)

Routine

Microscopic

Hepatic pedicle (hepatoduodenal ligament, retropancreatic and coeliac axis)

Adjuvant systemic

All

Ishida et al. [29] 2011

43

12 (27.9 %)

Routine

Both

Common hepatic, proper hepatic artery, superior border of pancreas

Adjuvant local or systemic

All

Rosen et al. [20] 1992

40

9 (22.5 %)

Selective

Both

Not stated

Not stated

Not stated

Aoki et al. [25] 2008

187

9 (4.8 %)

Selective

Both

Hepatoduodenal ligament

Adjuvant local or systemic or none

Not stated

Jaeck et al. [3] 2002

160

17 (10.6 %)

Routine

Both

Hepatic pedicle (hepatoduodenal, retropancreatic, coeliac axis, hepatic artery)

Not stated

Not stated

The overall prevalence of positive lymph nodes in the included studies was 8.4 %. The prevalence was 15.7 % in patients who underwent routine lymphadenectomy and it was 4.8 % in patients who underwent selective lymphadenectomy. The prevalence ranged between 2.6 and 50 %.

The overall 3-year and 5-year survival rates in node positive patients were 21.8 % and 8.5 %, respectively, compared with 58.2 % and 47.5 % in node negative patients (Table 2).
Table 2

Survival in different studies

Authors and year of publication

Total number of patients followed up for outcome

Node positive patients followed up

Node negative patients followed up

Total

1-year survival

3-year survival

5-year survival

Total

1-year survival

3-year survival

5-year survival

Ambiru et al. [17] 1999

149

8

 

1 (12.5 %)

1 (12.5 %)

141

 

63 (44.7 %)

38 (27.0 %)

Minagawa et al. [24] 2000

235

6

0

0

0

229

119 (52.0 %)

89 (38.9 %)

62 (27.1 %)

Carpizo et al. [9•] 2009

127

27

 

8 (29.6 %)

3 (11.1 %)

100

 

67 (67 %)

49 (49 %)

Yasui et al. [22] 1995

52

8

 

2 (25 %)

0 (0 %)

44

 

26 (59.1 %)

18 (40.9 %)

Jonas et al. [27] 2007

204

27

 

7 (25.9 %)

 

177

 

96 (54.2 %)

 

Settmacher et al. [28] 2011

382

18

   

364

   

Lupinacci et al. [30] 2013

26

5

   

21

   

Oussoultzoglou et al. [26] 2008

132

12

 

4 (33.3 %)

0 (0 %)

120

 

88 (73.3 %)

48 (40.0 %)

Adam et al. [7•] 2008

757

47

 

18 (38 %)

8 (18 %)

710

 

483 (68 %)

376 (53 %)

Beckurts et al. [4] 1997

119

35

 

1 (2.9 %)

0 (0 %)

84

 

40 (47.6 %)

19 (22.6 %)

Pulitanò et al. [8•] 2012

1519

61

52 (85.2 %)

21 (34.4 %)

11 (18.0 %)

1458

  

831 (57.2 %)

Harms et al. [21] 1999

155

39

 

1 (2.6 %)

0 (0 %)

116

 

52 (44.8 %)

24 (20.7 %)

Laurent et al. [5] 2004

156

23

16 (69.6 %)

6 (26.1 %)

1 (4.3 %)

133

117 (88 %)

74 (55.6 %)

57 (42.9 %)

Bennett et al. [11] 2008

45

8

 

2 (25 %)

 

37

 

28 (75 %)

 

Kokudo et al. [18] 1998

94

7

6 (85.7 %)

2 (28.6 %)

0 (0 %)

87

84 (96.6 %)

57 (65.5 %)

47 (54.0 %)

Sanchez-Cespedes et al. [23] 1999

15

8

2 (25 %)

0 (0 %)

 

7

7 (100)

2 of 4 (50 %)

 

Nakamura et al. [19] 1992

21

5

 

2 (40 %)

 

16

 

10 (62.5 %)

 

Ishida et al. [29] 2011

43

12

   

31

   

Rosen et al. [20] 1992

40

9

 

1 (11.1 %)

0 (0 %)

31

 

17 (54.8 %)

11 (35.5 %)

Aoki et al. [25] 2008

187

9

 

5 (55.6 %)

0 (0 %)

178

 

87 (48.9 %)

55 (30.9 %)

Jaeck et al. [3] 2002

160

17

3 (17.6 %)

3 (17.6 %)

0 (0 %)

143

135 (94.4 %)

89 (62.2 %)

67 (46.9 %)

The overall mortality was significantly higher in node positive patients than in those who were node negative (HR 2.72, 95 % CI 1.95–3.80) (Fig. 1). The odds of mortality at 1 year (OR 13.52, 95 % CI 2.70–67.70), 3 years (OR 4.54, 95 % CI 3.15–6.54) and 5 years (OR 6.33, 95 % CI 4.28–9.36) were significantly higher in the node positive group than in the node negative groups. In the subgroup of studies which included only patients with microscopic nodal involvement, mortality was significantly higher for node positive patients than for node negative patients (HR 2.79, 95 % CI 2.73–2.85). In the subgroup of studies which included only patients in whom routine hepatic lymphadenectomy was performed, mortality was higher for patients with node positive disease (HR 2.94, 95 % CI 1.59–5.43). In the subgroups classified by the extent/location of lymph node dissection, the mortality was significantly higher in patients who underwent clearance of hilar or perihepatic lymph nodes only (HR 2.08, 95 % CI 2.05–2.10), clearance of hepatoduodenal lymph nodes only (HR 2.68, 95 % CI 2.20–3.25) or clearance of hepatic pedicle lymph nodes only (HR 2.77, 95 % CI 1.12–6.83) compared with node negative patients. In studies where adjuvant or neoadjuvant chemotherapy was used, mortality was significantly higher for node positive patients (HR 2.42, 95 % CI 2.15–2.73).
https://static-content.springer.com/image/art%3A10.1007%2Fs11888-013-0165-6/MediaObjects/11888_2013_165_Fig1_HTML.gif
Fig. 1

Forest plot and funnel plot for overall mortality

Although there was significant heterogeneity in the main results and many of the subgroup analyses, the heterogeneity was in the magnitude of the effect rather than the direction. Visual assessment of the funnel plot did not demonstrate reporting bias (Fig. 1).

Morbidity data directly related to lymphadenectomy were not available for meta-analysis.

Discussion

The overall prevalence of positive lymph nodes was 8.4 %. The prevalence ranged between 2.6 and 50 %. Twelve studies [35, 11, 17, 19, 21, 22, 26, 27, 29, 30] reporting routine lymphadenectomy had a prevalence of positive lymph nodes ranging between 5.4 and 27.9 %, and five studies [5, 11, 19, 23, 30] which reported selective lymphadenectomy had a prevalence of positive lymph nodes ranging between 4.0 and 22.5 %. The prevalence of positive lymph nodes in the eight studies [4, 17, 18, 21, 24, 25, 27, 30] which reported hepatoduodenal/hepatic hilum lymph node dissection ranged between 2.6 and 27.8 %, and in the ten studies [3, 5, 7•, 8•, 9•, 11, 19, 22, 26, 29] which reported more extensive lymph node dissection of the hepatic pedicle and retropancreatic area, the prevalence ranged between 4.0 and 27.9 %. Thus, neither the extent of the lymphadenectomy nor the routine/selective approach for lymphadenectomy appeared to be the main factor causing the difference in prevalence of positive lymph nodes. This suggests that the wide variation in prevalence is related to patient selection or the histological technique used to detect lymph node metastases.

Two studies found increased sensitivity for detection of micrometastases in perihepatic lymph nodes by adding immunohistochemistry analysis [11] and molecular-based techniques [23]. These techniques could influence the prevalence of hepatic lymph node metastases, but larger studies are needed to assess their value and relevance in routine clinical practice. Exclusion of involved macroscopic lymph nodes [5], preoperative and intraoperative assessment, the number of nodes examined and neoadjuvant chemotherapy are other factors which may influence the prevalence of node positive disease.

The meta-analysis has clearly demonstrated that the long-term survival was considerably higher in those undergoing liver resection with no involved hepatic lymph nodes than in those with hepatic lymph node metastases. Survival was reduced regardless of whether the nodal involvement was microscopic or macroscopic, and whether the nodal involvement was found at the time of routine or selective lymphadenectomy.

Five studies [5, 11, 19, 23, 30] looked at survival in patients who had undergone a liver resection and lymph node dissection, and were found on subsequent histology to have microscopic hepatic lymph node involvement. Two of the five studies [5, 30] excluded patients with macroscopically involved nodes and two others [11, 19] sampled suspicious or palpable nodes. The overall mortality was significantly higher in the microscopically involved node positive patients than in the node negative patients. The studies which included patients with macroscopic node involvement also had patients with microscopic disease detected on intraoperative sampling of suspicious nodes. The survival data on patients with macroscopic disease were not available separately for meta-analysis to determine if there is a survival difference between macroscopic and microscopic lymph node involvement. However, studies have independently reported significant association with worse survival with intraoperative detection of nodal disease [7•, 9•], and this paradox needs to be explored further.

Data on the pattern of recurrent disease were not available separately for the meta-analysis. It would have been important to know if the patterns of recurrence differ and whether recurrences were mainly nodal in those with resection of nodal metastases

Surgeons usually associate mortality rates with surgical or operative mortality. Data on morbidity directly related to hepatic lymphadenectomy were not separately available for meta-analysis. Complications such as portal vein injury, duodenal injury, lymphatic leakage and bleeding have been associated with hepatic lymphadenectomy [31]. None of the included studies reported complications directly related to hepatic lymphadenectomy. Only two studies [3, 7•] compared morbidity in those who underwent lymphadenectomy with those who did not, and they reported no difference in morbidity between the two groups. None of the included studies reported higher mortality with lymphadenectomy. Generally, lymphadenectomy is a safe procedure in experienced hands but it is associated with potential risks [31].

There have been no randomized controlled trials comparing liver resection alone with liver resection along with routine lymphadenectomy, and hence there is no high-level evidence regarding the benefits of or harm from routine lymphadenectomy. For patients who are found to have macroscopic evidence of hepatic lymph node metastases at the time of surgery, there are no trials which have compared lymphadenectomy along with liver resection with systemic chemotherapy. Routine lymphadenectomy is generally performed to remove the microscopically involved nodes draining the liver and provide cancer clearance. The studies which involved routine lymphadenectomy [35, 11, 17, 19, 21, 22, 26, 27, 29, 30] reported very poor or no 5-year survival in these patients compared with node negative patients. There are several questions related to surgical aspects of lymphadenectomy which cannot be answered in this review on account of the small number of patients in the included series such as the following: Was the lymphadenectomy adequate? Was there variation in technique and node harvest? Does lymphadenectomy not work or is it being done poorly? Is there a learning curve?.

Perihepatic node metastases identified by haematoxylin and eosin staining or immunohistochemistry had a strong association with cancer recurrence [11]. None of the included studies compared the survival or recurrence rates between those who underwent routine lymphadenectomy and those who underwent selective lymphadenectomy (based on finding macroscopic disease). This would have been important information because the pattern of recurrence would help identify the reason for relapse, such as more extensive disease, more liver metastases or more extensive lymphatic dissemination.

Two studies [25, 30] reported no statistically significant survival difference between hilar node positive patients and node negative patients. In the study by Aoki et al. [25], there was no significant difference in the proportion of patients alive at 3 years between the node positive and node negative groups (HR 1.58, 95 % CI 0.64–2.52). However the number of node positive patients in this study was small (n = 9), and there were no 5-year survivors. The other study, by Lupinacci et al. [30], had five node positive patients whose median survival and recurrence-free survival were similar to those of the node negative groups. The small numbers are a considerable limitation of these studies and it is difficult to draw meaningful conclusions about the observed lack of difference in survival.

In the study by Adam et al. [7•], nine patients (of a cohort of 47 patients) who underwent combined liver resection and pedicular lypmphadenectomy, and were under 40 years of age and had responded to aggressive preoperative chemotherapy, had a 5-year survival rate of 45 %. Patients who did not respond to chemotherapy were excluded and patients over 40 years of age had a 5-year survival rate of only 10 %. Whether age is a factor in outcome is not clear from this series. In the study by Jaeck et al. [3], of the 17 patients who had undergone liver resection and routine hepatic pedicle lymphadenectomy, 19 % survived for 3 years. Pulitanò et al. [8•] found 3-year and 5-year overall survival rates of 35 % and 18 %, respectively, in 61 patients with node positive disease, most of whom had received preoperative as well as adjuvant chemotherapy. These studies demonstrate that a small number of highly selected patients may benefit from surgical resection to achieve 3-year and 5-year overall survival in an attempt to shift the goal from palliation to prolonging survival. Although, by virtue of the study design and small numbers, the results cannot be extrapolated generally, these studies could become a stimulus for larger studies with stratified groups which are needed to draw meaningful conclusions.

Three studies [3, 7•, 8•] compared survival of patients with positive nodes in different locations and all three found better outcome in patients with hepatoduodenal lymph node involvement as against those with coeliac and para-aortic lymph node involvement. Importantly, Adam et al. [7•] and Pulitanò et al. [8•] found exceptionally poor survival in patients with coeliac and para-aortic node involvement despite preoperative chemotherapy. Again, despite the limitations of the study design and small numbers, it is a relevant finding that extensive lymphadenectomy did not confer benefit. Traditionally, involvement of coeliac and para-aortic lymph nodes has been an absolute contraindication to surgery. There is currently no evidence to support the resection of colorectal liver metastases along with a coeliac or para-aortic lymphadenectomy.

Fifteen studies [5, 7•, 8•, 9•, 11, 1719, 21, 2530] included patients who received preoperative and/or postoperative chemotherapy. In the meta-analysis, the overall odds of mortality were significantly higher in the node positive patients where adjuvant or neoadjuvant chemotherapy was used (HR 2.42, 95 % CI 2.15–2.73). In many of these studies all the patients received chemotherapy and the studies did not analyse stratified patient groups. Aoki et al. [25] suggested that favourable survival in four patients in their series may be related to a newer regimen of postoperative chemotherapy. Adam et al. [7•] commented that the 5-year survival rate of 18 % in node positive patients may be related to preoperative chemotherapy and patient selection based on absence of progression of nodal disease before surgery. The survival benefit was seen in patients with pedicular lymph node involvement and not in patients with coeliac and para-aortic node involvement. Similarly, Carpizo et al. [9•] found a 5-year survival rate of only 12 % in portal node positive patients and no 5-year survivors with positive retroperitoneal nodes. Important information such as the outcome of adjuvant therapy guided by positive lymph nodes is not available from these studies. It is possible that modern chemotherapy has an impact on survival in hepatic pedicular lymph node positive patients, but this needs to be demonstrated in large and properly designed clinical trials. Until such data are available, clinicians offering liver resection along with hepatic lymphadenectomy for resectable liver metastases associated with hepatic lymph node metastases must inform their patients that the surgery is without evidence of a survival advantage. There were no randomized controlled trials comparing liver resection versus other treatments in patients with colorectal cancer liver metastases with lymph node involvement.

Conclusions

We conclude that there is no evidence of survival benefit from routine lymphadenectomy in patients with resectable colorectal liver metastases. In patients with hepatic nodal metastases as well as liver metastases, liver resection and lymphadenectomy is of no proven benefit and has not been compared with systemic chemotherapy. Survival rates are low in patients with positive lymph nodes draining the liver irrespective of whether they are detected by routine lymphadenectomy or by macroscopic involvement. Our practice is to counsel patients in whom liver resection along with lymphadenectomy is being considered that the procedure is not part of routine clinical practice and is unlikely to be curative, but it may prolong survival. The procedure would ideally be performed as part of a clinical study or trial. Patients with colorectal liver metastases and hepatic nodal involvement could be considered for a trial offering surgical resection (liver resection and lymphadenectomy) and chemotherapy versus chemotherapy alone for patients with prior response to systemic chemotherapy.

Conflict of Interest

Rahul S. Koti declares he has no conflict of interest.

Constantinos Simillis declares he has no conflict of interest.

Kurinchi S. Gurusamy is supported by grants from the Wellcome Trust and the National Institute for Health Research (NIHR), and he has also received payment for preparation of a manuscript from NIHR.

Michael Jacovides declares he has no conflict of interest.

Brian R. Davidson declares he has no conflict of interest.

Copyright information

© Springer Science+Business Media New York 2013