Available online at www.sciencedirect.

com

EJSO 34 (2008) 55e60 www.ejso.com

Elevated preoperative neutrophil to lymphocyte ratio predicts

survival following hepatic resection for colorectal liver metastases
K.J. Halazun, A. Aldoori, H.Z. Malik, A. Al-Mukhtar, K.R. Prasad,
G.J. Toogood, J.P.A. Lodge*
HPB and Transplant Unit, St. James’s University Hospital, Leeds LS9 7TF, UK
Accepted 7 February 2007
Available online 19 April 2007

Abstract
Background: The neutrophilelymphocyte ratio (NLR) provides an indicator of inflammatory status. An elevated NLR has been shown to be
a prognostic indicator in primary colorectal malignancy. The aim of this study was to establish whether NLR predicts outcome in patients
undergoing resection for colorectal liver metastasis.
Design: Retrospective analysis of the white cell and differential counts for 440 patients undergoing liver resections for colorectal liver
metastasis between January 1996 and January 2006. An NLR  5 was considered to be elevated.
Results: Two hundred and eighty-nine males and 151 females were included. Seventy-eight patients (18%) had an elevated NLR, 55 of
whom died, giving elevated NLR a positive predictive value (PPV) for death of 71%. Sixty of the 78 patients had recurrent disease giving
raised NLR an PPV for recurrence of 78%. The 5-year survival for patients undergoing resection with high NLR was significantly worse
than that for patients with normal NLR (22% vs. 43%, p < 0.0001). Univariate analysis of factors affecting survival revealed raised NLR,
number of metastases >8, tumour size >5 cm and age >70 significantly affected outcome. All factors except tumour size remained sig-
nificant predictors of term survival on multivariate analysis (NLR:HR ¼ 2.261, CI ¼ 1.654e3.129, p < 0.0001, metastases >8:HR ¼ 1.611,
CI ¼ 1.006e2.579, p ¼ 0.047, age >70:HR ¼ 1.418, CI ¼ 1.049e1.930, p ¼ 0.027). Elevated NLR was found to be the sole positive pre-
dictor of recurrence on univariate analysis (HR ¼ 4.521, CI ¼ 2.475e8.257, p < 0.0001).
Conclusion: Elevated NLR increases both risk of death and the risk of recurrence in patients who undergo surgery for CRLM. Preoperative
NLR measurement may therefore provide a simple method of identifying patients with a poorer prognosis.
Ó 2007 Elsevier Ltd. All rights reserved.

Keywords: Colorectal liver metastases; Neutrophil to lymphocyte ratio; Prognosis

Introduction Several studies have searched for prognostic indicators

Liver resection is the primary mode of treatment for
patients with CRLM, offering the only potential for disease
eradication and therefore cure. Five-year survival approaches
45e50% in some centres,1e6 however, cure from CRLM
after surgery only occurs in 15e20% of patients. In addition,
both intra- and extra-hepatic tumour recurrence rates remain
high at around 60e65%.7,8 The identification of patients
more likely to have recurrence or poor outcome after surgery
is therefore important and useful in guiding treatment.
* Corresponding author. Tel.: þ44 (0) 113 2064890; fax: þ44 (0) 113
2448182.
E-mail address: peter.lodge@leedsth.nhs.uk (J.P.A. Lodge).
of outcome for CRLM patients. Potential prognostic indica-
tors include primary tumour stage and grade, size, distribu-
tion and number of liver metastases, extra-hepatic disease,
resection margins and lymph node status.4,9e13 Although
these histological and surgical prognostic indicators are
valuable, they have not been widely applied and little exists
by way of a consensus for selecting patients who would
benefit most from surgery and adjuvant chemotherapy.
More recently, there has been growing interest in the host’s
inflammatory response to tumour, and the systemic effects
exerted by tumours in causing upregulation of the inflam-
matory process, thereby increasing propensity to metasta-
sise through the inhibition of apoptosis, promotion of
angiogenesis and damage of DNA.14e18

0748-7983/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ejso.2007.02.014
56 K.J. Halazun et al. / EJSO 34 (2008) 55e60
The most widely studied measure of inflammation is
C-reactive protein (CRP), levels of which have been shown
to independently predict survival in patients who undergo
curative resection for colorectal cancer.19,20 Recently, our
group has identified CRP as a prognostic indictor in pa-
tients undergoing surgery for CRLM.21 A further marker
of inflammation that is increasingly used to assess outcome
in critically ill surgical patients is the neutrophil to lympho-
cyte ratio (NLR). An elevated NLR has been shown to be
an indictor of poor outcome in vascular and cardiovascular
patients undergoing intervention.22,23 Walsh et al.24 have
also shown an NLR  5 to be a marker of survival in colo-
rectal cancer patients. We therefore hypothesise that an el-
evated NLR may be used as a preoperative prognostic
indictor of both outcome and recurrence in CRLM patients
undergoing curative hepatic resection.
Patients and methods
Calculation of NLR
Patients undergoing resection for colorectal liver metasta-
ses had neutrophil and lymphocyte counts measured preoper-
atively as part of the routine work up. All white cell and
differential counts were taken on the day before surgery
with none of the patients showing clinical signs of sepsis.
The NLR was calculated from the differential count by divid-
ing the neutrophil measurement by the lymphocyte measure-
ment. An NLR  5 was considered elevated. Patients were
excluded if preoperative full blood counts were unavailable
or the surgery was synchronous with primary colorectal
resection.
Patient selection and surgical technique
The criteria for acceptance for surgery included fitness
for major resection and lack of disseminated or irresectable
extra-hepatic disease identified by CT or MRI scan. In all
cases studied the colorectal primary had been previously
resected and the patients had recovered fully from that pro-
cedure. Intra-operative ultrasound was used as an adjunct to
the preoperative radiological investigations. Resection was
performed using the Cavi-Pulse Ultrasonic Surgical Aspira-
tor (CUSA, Model 200T, Valley Lab., Boulder Colorado,
USA). If necessary an intermittent Pringle manoeuvre
was used with 15 min of ischaemia followed by 5 min of
reperfusion. All patients undergoing liver resection accord-
ing to our unit protocol were offered adjuvant therapy in the
form of 5-FU/folinic acid, unless they had received adju-
vant therapy following their colonic resection within the
past 1 year. Patients were followed up at specialist clinics,
with a minimum follow-up period of 11 months at the time
of writing (range 11e97 months; median 24 months). No
patients were lost to follow-up. An intensive policy of
post-operative surveillance exists within this unit. Patients
have 3 monthly chest and abdominal computerised tomo-
graphy (CT) performed during the 1st post-operative year,
then 6 monthly during year 2. From years 3e5, a CT
scan is performed yearly and finally at years 7 and 10 of
follow-up. Tumour markers (CEA, CA19-9) and liver func-
tion tests are performed during each clinic visit. The data
examined included patient demographics; liver resection
histology; pre-hepatectomy NLR; post-operative morbid-
ity/mortality results as well as recurrence and survival
figures.
Statistical methods
SPSS version 11 and Graph Pad Prism 4 for Mac were
used to analyse the data. Chi-squared and Fisher’s Exact
tests were used to analyse differences among groups of pa-
tients with high or normal NLR. Kaplan-Meier survival
curves were used to analyse patient outcome and disease
free survival. A Cox regression analysis was then per-
formed in a step-wise manner in order to perform a multi-
variable analysis of clinico-pathological factors that impact
both overall survival. Binary logistic regression analysis
was used to investigate factors influencing recurrence. All
confidence intervals are 95%.
Results
A total of 440 patients were included in this study. Of
these patients 289 (65%) were males and 151 (35%)
females. The mean age of patient at time of surgery was
64 years (range 32e88 years; S.D. 10.7 years). All patients
underwent liver resection. A total of 266 patients (61%) had
a ‘‘major’’ (three or more Couinaud’s segments) resection
performed. The in-hospital mortality rate was 2.5%, overall
(long term) mortality was 42% and 52% of patients devel-
oped recurrence.
Predictive value of NLR
The preoperative NLR was elevated (5) in 78 patients
(18%). Of these 55 patients have died, therefore giving
elevated NLR a positive predictive value for death of
71%. Recurrence occurred in 60 of the 78 patients with
a high NLR, giving high NLR a positive predictive value
for recurrence of 78%.
Elevated NLR effects survival and recurrence
There was no significant difference in demographic
and clinico-pathological features between patients with
normal and high NLR (Table 1). There was a significant
difference in long term survival between patients with
a normal NLR and those with a high NLR as shown in
Fig. 1. Five-year survival for patients with a normal
NLR was 43% compared to 22% for patients with a raised
NLR (log rank test p < 0.0001). A marked difference also
 K.J. Halazun et al. / EJSO 34 (2008) 55e60 57

Table 1 100
Comparison of clinico-pathological and demographic features
Factor Normal Elevated Significance Normal NLR
NLR (<5) NLR (5) ( p) Elevated NLR
N ¼ 362 N ¼ 78 75

Age (mean) 64.1 64.02 0.419

Male gender 234 55 0.291

% Survival
Neoadjuvant 41 6 0.424
chemotherapy 50
Synchronous disease 121 25 0.554
Multiple (eight or 41 11 0.441
more) metastases
Major resections 217 49 0.608 25
(>3 segments)
Large metastases size 137 37 0.122
(>50 mm)
Log Rank Test: p < 0.0001
Positive margin 126 27 0.893
0
0 1 2 3 4 5
Disease Free Survival - Years

existed between the two groups of patients with respect to No. at Risk (% survival)
disease free survival; patients with a high NLR having a 5 Year 0 1 2 3 4 5

year disease free survival of 12% compared to 42% in Normal NLR 362 259 (72%) 119 (56%) 70 (49%) 42 (43%) 24 (42%)

patients with a normal NLR (log rank test p < 0.0001) High NLR 78 39 (57%) 15 (26%) 11 (20%) 7 (14%) 5 (12%)

(Fig. 2). Univariate analysis of factors affecting overall out-

come (long term survival) revealed that tumour size >5 cm, Figure 2. KaplaneMeier chart showing disease free survival in patients
tumour number (>8), Age (>70) and elevated NLR (5) with normal and high NLR.
influenced overall survival (Table 2). Elevated NLR,
age and tumour number remained significant on multi- Discussion
variate analysis (Table 3). Elevated NLR was found to
be the sole predictor of recurrence on univariate analysis Inflammation, elevated NLR and malignancy
(Table 4).
The first casual link between cancer and inflammation
was described over one and a half centuries ago by Rudolf

100
Table 2
Univariate analysis of factors affecting overall survival
Normal NLR
p Hazard ratio (CI)
Elevated NLR
Age
75 <70 (n ¼ 303)
>70 (n ¼ 137) 0.037 1.381 (1.020e1.869)
Gender
% Survival

F (n ¼ 151)
50 M (n ¼ 289) 0.494 1.110 (0.824e1.494)
No. of tumours
<8 (n ¼ 388)
>8 (n ¼ 52) 0.05 1.535 (0.997e2.364)
Size of largest tumour
25
<5 cm (n ¼ 249)
>5 cm (n ¼ 174) 0.011 1.464 (1.039e1.962)
Timing of tumour
Log Rank Test: p < 0.0001
Synchronous (n ¼ 146)
0 Metachronous (n ¼ 153) 0.891 1.023 (0.738e1.418)
0 1 2 3 4 5
No. of segments removed
Survival Years <3 (n ¼ 174)
>3 (n ¼ 266) 0.341 1.156 (0.858e1.559)
No. at Risk (% survival)
Preoperative NLR
Year 0 1 2 3 4 5
<5 (n ¼ 362)
Normal NLR 362 322 (90%) 187 (79%) 117 (67%) 73 (57%) 36 (43%) 5 (n ¼ 78) <0.0001 2.261 (1.644e3.110)
High NLR 78 54 (71%) 35 (53%) 21 (41%) 11 (25%) 9 (22%) Resection margin
ve (n ¼ 269)
+ve (n ¼ 153) 0.390 1.141 (0.845e1.541)
Figure 1. KaplaneMeier chart comparing survival in both patient groups.
58 K.J. Halazun et al. / EJSO 34 (2008) 55e60

Table 3 patients with CRLM, as evidenced by a raised NLR, is

Multivariate analysis of factors affecting overall survival associated with poor overall and disease free survival, and
p Hazard ratio (CI) an increased risk of recurrence. This is consistent with the
Age > 70 0.027 1.418 (1.049e1.930) above studies which associate CRP and high NLR with poor
Tumour no. > 8 0.047 1.611 (1.006e2.579) outcome in primary colorectal tumours, as well as a study
NLR > 5 <0.0001 2.275 (1.654e3.129) from our unit which correlates high CRP levels with poor
Tumour size > 5 0.067 1.323 (0.981e1.784)
outcome in CRLM patients.16e21,24 This is the first study
to link elevated NLR and outcome after resection of CRLM:
Virchow, when he observed that leucocytes existed in neo- this study implicates elevated NLR with increased recur-
plastic tissue.14 It is only in the past decade, however, that rence risk and reduced disease free survival.
the complexities of the tumour inflammatory microenviron- The association between elevated NLR and poor progno-
ment, and the host’s response to tumour induced inflamma- sis is probably complex and largely unclear, however, several
tory pathways are beginning to be understood, resulting in possible explanations exist. The host’s immune response
an improved ability to prevent and treat malignancy. to tumour is lymphocyte dependent. Several studies have
Inflammation has been shown to play an important role demonstrated that patients with weaker lymphocytic infil-
in the pathogenesis and progression of colorectal carci- trates at tumour margins have a worse prognosis.27e30 Okano
noma. Links have been established through the greatly in- et al.30 found that patients with CRLM and weak lym-
creased risk of malignancy that exists in patients with phocytic infiltration at the tumour margin did worse after
inflammatory bowel disease,25 as well as the approximate liver resection than those with an adequate lymphocyte
50% decrease in colorectal cancer risk in patients who response to tumour. Patients with elevated NLR have a rela-
take regular NSAIDs and aspirin.26 In addition, elevated tive lymphocytopaenia, and, as a result may exhibit a poorer
markers of inflammation, especially elevated CRP, have lymphocyte mediated immune response to malignancy,
been used as prognostic tools in patients undergoing cura- thereby worsening their prognosis, and increasing the poten-
tive resection for primary colorectal tumours.16e20 A fur- tial for the tumour to recur.
ther preoperative marker, an elevated neutrophil to Alternatively, an elevated neutrophil count may aid in
lymphocyte ratio (NLR), has also been linked with poor the development and progression of the neoplasm by pro-
prognosis in patients with primary colorectal carcinoma.24 viding an adequate environment for it to grow. Circulating
Despite the frequency of metastasis of primary colorectal neutrophils have been shown to contain and secrete the vast
tumours to the liver, few inflammatory marker have been majority of circulating VEGF, a pro-angiogenic factor that
linked with prognosis in patients with CRLM. Our study is thought to play an integral role in tumour development.31
demonstrates that the preoperative inflammatory status of Increased angiogenic activity in GI tumours has been asso-
ciated with poor outcomes,32,33 The high circulating neu-
Table 4 trophil levels in patients with an elevated NLR may
Univariate analysis of factors affecting recurrence confer a survival advantage for metastatic colorectal
p Hazard ratio (CI) tumour cells, thus accounting for the poorer outcome and
increased recurrence rates in these patients.
Age
<70 (n ¼ 303)
>70 (n ¼ 137) 0.249 0.787 (0.523e1.183) Clinical use of elevated NLR
Gender
F (n ¼ 151) The ability to successfully predict poor prognosis and
M (n ¼ 289) 0.369 1.199 (0.807e1.783)
increased risk of recurrence in CRLM patients using NLR
No. of tumours
<8 (n ¼ 388) is potentially valuable in directing both pre- and post-
>8 (n ¼ 52) 0.348 1.328 (0.734e2.402) operative therapies to such patients in order to improve their
Size of largest tumour outcome. No specific therapies for these patients exist at
<10 cm (n ¼ 407) present, however, there has been ongoing research into
>10 cm (n ¼ 33) 0.343 1.421 (0.688e2.934)
the effects of anti-inflammatory agents on tumour progres-
Timing of tumour
Synchronous (n ¼ 146) sion. Several studies have shown that the anti-angiogenic
Metachronous (n ¼ 153) 0.680 0.906 (0.568e1.447) activity of COX-2 inhibitors, via the suppression of
No. of segments removed VEGF, inhibits and may even prevent the growth and pro-
<3 (n ¼ 174) liferation of CRLM.34e36 Such therapeutic strategies may
>3 (n ¼ 266) 0.237 1.262 (0.858e1.856)
prove most beneficial in patients with an elevated NLR in
Preoperative NLR
<5 (n ¼ 362) whom high levels of VEGF and other pro-angiogenic fac-
5 (n ¼ 78) <0.0001 4.521 (2.475e8.257) tors could be inhibited both pre- and post-operatively. Other
Resection margin studies have used therapies directed at enhancing the host
ve (n ¼ 269) response to tumour in order to decrease the propensity of
+ve (n ¼ 153) 0.092 1.414 (0.944e2.117)
colorectal tumours to propagate and metastasise.37,38 One
 K.J. Halazun et al. / EJSO 34 (2008) 55e60
such study by Oosterling et al. showed that the preoperative
administration of GM-CSF in patients undergoing surgery
for colorectal tumours enhances the immune activity of
the liver by increasing the hepatic population of dendritic
cells, and enhancing their interaction with CD8þ T lym-
phocytes.38 The possibility of preoperatively treating
patients with an elevated NLR with GM-CSF has not
been explored, but may improve their immune response to tu-
mour, and therefore decrease the risk of recurrence and im-
prove their long term survival. The use of vaccines to
enhance the lymphocyte response to tumour has also been
shown to be effective in colorectal cancer,39,40 and may again
be an avenue for treatment of patients with a high NLR.
Conclusion
In summary, elevated preoperative NLR increases both
risk of death and the risk of recurrence in patients who un-
dergo surgery for colorectal liver metastases. Preoperative
NLR measurement in such patients may provide a simple
method of identifying patients with a poorer prognosis
and aid in guiding treatment effectively. Although no spe-
cific therapy for such patients exists at present, pre- and
post-operative inflammatory and immune modulation may
prove beneficial in improving their long term outcome.
Acknowledgements
The authors would like to acknowledge Mr. Anthony
Kaye (Department of Haematology) for providing access
to patients’ blood results.
References
1. Jatzko GR, Lisborg PH, Stettner HM, Klimpfinger MH. Hepatic resec-
tion for metastases from colorectal carcinoma e a survival analysis.
Eur J Cancer 1995;31:41–6.
2. Fegiz G, Ramacciato G, Gennari L, et al. Hepatic resections for colo-
rectal metastases: the Italian multicenter experience. J Surg Oncol
1991;2:S144–54.
3. Simmonds PC, Primrose JN, Colquitt JL, Garden OJ, Poston GJ,
Rees M. Surgical resection of hepatic metastases from colorectal can-
cer: a systematic review of published studies. Br J Cancer 2006;94:
982–99.
4. Finch RJB, Malik HZ, Hamady ZZR, et al. Outcome following liver
resection for colorectal liver metastases: effect of type of resection.
Br J Surg, in press.
5. Taylor M, Forster J, Langer B, Taylor BR, Greig PD, Mahut C. A
study of prognostic factors for hepatic resection for colorectal metas-
tases. Am J Surg 1997;173:467–71.
6. Yamamoto J, Shimada K, Kosuge T, Yamasaki S, Sakamoto M,
Fukuda H. Factors influencing survival of patients undergoing hepa-
tectomy for colorectal metastases. Br J Surg 1999;86:332–7.
7. Hughes KS, Simon R, Songhorabodi S, et al. Resection of the liver for
colorectal carcinoma metastases: a multiinstitutional study of patterns
of recurrence. Surgery 1986;100:2782–4.
8. Nordlinger B, Quilichini MA, Parc R, Hannoun L, Delva E, Huguet C.
Hepatic resection for colorectal liver metastases. Influence on survival
59
of preoperative factors and surgery for recurrences in 80 patients. Ann
Surg 1987;205:256–63.
9. Neal CP, Garcea G, Doucas H, et al. Molecular prognostic markers in
resectable colorectal liver metastases: a systematic review. Eur J Can-
cer 2006;42(12):1728–43.
10. Takahashi S, Inoue K, Konishi M, Nakagouri T, Kinoshita T. Prognos-
tic factors for poor survival after repeat hepatectomy in patients with
colorectal liver metastases. Surgery 2003;133:627–34.
11. Tanaka K, Shimada H, Fujii Y, et al. Pre-hepatectomy prognostic stag-
ing to determine treatment strategy for colorectal cancer metastases to
the liver. Arch Surg 2004;389:371–9.
12. Fong Y, Fortner JG, Sun R, Brennan MF, Blumgart LH. Clinical score
for predicting recurrence after hepatic resection for metastatic colorec-
tal cancer: analysis of 1001 consecutive cases. Ann Surg 1999;230:
309–21.
13. Seifert JK, Bottger TC, Weigel TF, Gonner U, Junginger T. Prognostic
factors following liver resection for hepatic metastases from colorectal
cancer. Hepatogastroenterology 2000;47(31):239–46.
14. Balkwill F, Mantovani A. Inflammation and cancer: back to Virchow?
Lancet 2001;357(5):539–45.
15. Coussens LM, Werb Z. Inflammation and cancer. Nature 2002;420:
860–7.
16. Gunter MJ, Stolzenberg-Solomon R, Cross AJ, et al. A prospective
study of serum C-reactive protein and colorectal cancer risk in men.
Cancer Res 2006;66(4):2483–7.
17. Jaiswal M, LaRusso NF, Burgart LJ, Gores GJ. Inflammatory cyto-
kines induce DNA damage and inhibit DNA repair in cholangiocarci-
noma cells by a nitric oxide-dependent mechanism. Cancer Res 2000;
60:184–90.
18. Jackson JR, Seed MP, Kircher CH, Willoughby DA, Winkler JD. The
codependence of angiogenesis and chronic inflammation. FASEB J
1997;11:457–65.
19. McMillan DC, Canna K, McArdle CS. Systemic inflammatory
response predicts survival following curative resection of colorectal
cancer. Br J Surg 2003;90:215–9.
20. McMillan DC, Wotherspoon HA, Fearon KC, Sturgeron C, Cooke TG,
McArdle CS. A prospective study of tumor recurrence and the acute-
phase response after apparently curative colorectal cancer surgery. Am
J Surg 1995;170:319–22.
21. Wong V, Malik HZ, Hamady ZZ, et al. C-reactive protein as a predictor
of prognosis following curative resection for colorectal liver metasta-
ses. Br J Cancer 2007 Jan 29;96(2):222–5.
22. Duffy BK, Gurm HS, Rajagopal V, Gupta R, Ellis SG, Bhatt DL. Use-
fulness of an elevated neutrophil to lymphocyte ratio in predicting
long-term mortality after percutaneous coronary intervention. Am
J Cardiol 2006 Apr 1;97(7):993–6.
23. Zahorec R. Ratio of neutrophil to lymphocyte counts e rapid and sim-
ple parameter of systemic inflammation and stress in critically ill. Bra-
tisl Lek Listy 2001;102:5–14.
24. Walsh SR, Cook EJ, Goulder F, Justin TA, Keeling NJ. Neutrophil-
lymphocyte ratio as a prognostic factor in colorectal cancer. J Surg
Oncol 2005 Sep 1;91(3):181–4.
25. Eaden JA, Abrams KR, Mayberry JF. The risk of colorectal cancer in
ulcerative colitis: a meta-analysis. Gut 2001;48:526–35.
26. IARC Working Group. IARC handbook of cancer prevention. In: Non-
steroidal anti-inflammatory drugs, vol. 1. Lyon, France: IARC-WHO;
1997.
27. Ropponen KM, Eskelinen MJ, Lipponen PK, et al. Prognostic value of
tumor-infiltrating lymphocytes (TILs) in colorectal cancer. J Pathol
1997;182:318–29.
28. Jass JR. Lymphocytic infiltration and survival in rectal cancer. J Clin
Pathol 1986;39:585–9.
29. Svennevig JL, Lunde OC, Holter J, et al. Lymphoid infiltration and
prognosis in colorectal carcinoma. Br J Cancer 1984;49:375–7.
30. Okano K, Maeba T, Modoguchi A, et al. Lymphocytic infiltration sur-
rounding liver metastases from colorectal cancer. J Surg Oncol 2003;
82(1):28–33.
60 K.J. Halazun et al. / EJSO 34 (2008) 55e60
31. Kusumanto YH, Dam WA, Hospers GA, Meijer C, Mulder NH. Plate-
lets and granulocytes, in particular the neutrophils, form important
compartments for circulating vascular endothelial growth factor. An-
giogenesis 2003;6(4):283–7.
32. Fondevila C, Metges JP, Fuster J, et al. p53 and VEGF expression
are independent predictors of tumour recurrence and survival fol-
lowing curative resection of gastric cancer. Br J Cancer 2004;90:
206–15.
33. Tanigawa N, Amaya N, Matsumura M, Shimomatsuya T. Correlation
between expression of vascular endothelial growth factor and tumor
vascularity, and patient outcome in human gastric carcinoma. J Clin
Oncol 1997;15:826–32.
34. Yamauchi T, Watanabe M, Hasegawa H, et al. The potential for a selec-
tive cyclooxygenase-2 inhibitor in the prevention of liver metastasis
in human colorectal cancer. Anticancer Res 2003 JaneFeb;23(1A):
245–9.
35. Fenwick SW, Toogood GJ, Lodge JP, Hull MA. The effect of the
selective cyclooxygenase-2 inhibitor rofecoxib on human colorectal
cancer liver metastases. Gastroenterology 2003 Sep;125(3):716–29.
36. Chalmers CR, Wilson DJ, Ward J, Robinson PJ, Toogood GJ,
Hull MA. Antiangiogenic activity of the selective cyclooxygenase 2
inhibitor rofecoxib in human colorectal cancer liver metastases. Gut
2006 Jul;55(7):1058–9.
37. Alves A, Vibert E, Trajcevski S, et al. Adjuvant interleukin-12 gene
therapy for the management of colorectal liver metastases. Cancer
Gene Ther 2004 Dec;11(12):782–9.
38. Oosterling SJ, Mels AK, Geijtenbeek TB, et al. Preoperative granulo-
cyte/macrophage colony-stimulating factor (GM-CSF) increases he-
patic dendritic cell numbers and clustering with lymphocytes in
colorectal cancer patients. Immunobiology 2006;211(6e8):641–9.
39. Pilla L, Squarcina P, Coppa J, et al. Natural killer and NK-Like T-cell
activation in colorectal carcinoma patients treated with autologous tu-
mor-derived heat shock protein 96. Cancer Res 2005 May 1;65(9):
3942–9.
40. Lesterhuis WJ, de Vries IJ, Schuurhuis DH, et al. Vaccination of co-
lorectal cancer patients with CEA-loaded dendritic cells: antigen-spe-
cific T cell responses in DTH skin tests. Ann Oncol 2006 Jun;17(6):
974–80.