original article Annals of Oncology 18: 1030–1036, 2007

doi:10.1093/annonc/mdm085
Published online 10 April 2007

Vascular endothelial growth factor gene

polymorphisms associated with prognosis for
patients with gastric cancer
J. G. Kim1*, S. K. Sohn1, Y. S. Chae1, Y. Y. Cho1, H.-I. Bae2, G. Yan2, J. Y. Park3,
M.-H. Lee4, H. Y. Chung5 & W. Yu5
1
Department of Oncology/Hematology, 2Department of Pathology, 3Department of Internal Medicine and Biochemistry, 4Department of the Advanced Medical
Technology Center for Diagnosis and Prediction, and 5Department of Surgery, Kyungpook National University Hospital, Kyungpook National University School of
Medicine, Daegu, Korea

Received 24 October 2006; revised 23 December 2006; accepted 8 February 2007

Background: The present study analyzed vascular endothelial growth factor (VEGF) gene polymorphisms and their
impact on the prognosis for patients with gastric cancer.
Patients and methods: Five hundred and three consecutive patients with surgically resected gastric
adenocarcinoma were enrolled in the present study. The genomic DNA was extracted from paraffin-embedded tissue
and four VEGF (2460T > C, 2116G > A, +405G > C, and +936C > T) gene polymorphisms were determined using
a polymerase chain reaction-restriction fragment length polymorphism assay.
Results: The survival analysis showed no association of three VEGF gene polymorphisms with the prognosis. For the
+936C > T polymorphism, the T/T genotype, however, had a worse overall survival (OS) compared with the C/C
original
article

genotype (P = 0.037). The 2460 T/C or C/C genotype was a poor prognostic factor in patients with stage 0 or I gastric
cancer (OS: hazard ratio (HR) = 3.96, disease-free survival (DFS): HR = 4.87). In the haplotype analysis, the
CACC haplotype was associated with a significantly worse survival when compared with the TGGC haplotype
(OS: HR = 1.72, DFS: HR = 1.73).
Conclusions: VEGF gene polymorphisms were found to be an independent prognostic marker for patients with
gastric cancer. Consequently, the analysis of VEGF gene polymorphisms can help identify patient subgroups at
high risk of a poor disease outcome.
Key words: gastric cancer, gene, polymorphism, prognosis, VEGF

introduction regulation of the lymphangiogenic process, has been reported

Although gastric cancer is increasingly being diagnosed at an
early stage, resulting in 10-year survival rates of between 80%
and 95% in the Far East, gastric adenocarcinoma is still the
second leading cause of cancer-related death worldwide [1, 2].
Angiogenesis, the formation of new blood vessels from
endothelial precursors, is a prerequisite for the growth and
progression of solid malignancies, and the vascular endothelial
growth factor (VEGF) superfamily of endothelial growth
factors has been identified to critically influence tumor-related
angiogenesis [3, 4]. Clinical studies have demonstrated that an
increased expression of VEGF or its family is associated with
the grade of angiogenesis and the prognosis for various human
cancers [5–8]. In particular, with gastric cancer, the expression
of VEGF or VEGF-C, which are intimately involved in the
*Correspondence to: Dr J. G. Kim, Department of Oncology/Hematology, Kyungpook
National University Hospital, Kyungpook National University School of Medicine,
50 Samduck 2-Ga, Jung-Gu, Daegu 700-712, South Korea. Tel: +82-53-420-6522;
Fax: +82-53-426-2046; E-mail: jkk21c@mail.knu.ac.kr
to be correlated with a poor prognosis [9–11]. Furthermore,
Juttner et al. [12] found that the presence of VEGF-D and
its receptor vascular endothelial growth factor receptor 3 was
associated with lymphatic metastasis, reduced patient survival,
and poor prognosis after the curative resection of gastric
adenocarcinomas. Given these results, VEGF or its family
would seem to play an important role in lymphangiogenesis
and lymphatic tumor spread, thereby affecting the prognosis
of gastric cancer.
The VEGF gene is assigned to chromosome 6p12-p21 and
consists of eight exons separated by seven introns that exhibit
alternative splicing to form a family of proteins [13, 14]. Several
polymorphisms have been described in the VEGF gene, and
some of these variants [2460T > C, 2116G > A, and +405G >
C (transcription start site counted as +1) in the promoter or
5#-untranslated region and +936C > T in the 3#-untranslated
region] found to be associated with variations in VEGF protein
production [15–17]. In clinical studies, these polymorphisms
have been reported to be involved in the development of solid

ª 2007 European Society for Medical Oncology

Annals of Oncology
tumors, such as melanoma [18], lung [19], prostate [20], and
breast cancer [21], where angiogenesis is critical in the
pathogenesis of the disease. Moreover, recent studies have
demonstrated that genetic polymorphisms can be used to
predict the clinical outcomes of gastric [22], breast [23, 24],
colon [25], and pancreatic cancer [26].
No study, however, has yet been published that has
investigated the single nucleotide polymorphisms (SNPs) of the
VEGF gene and their relationship to the clinical outcomes of
gastric cancer. Therefore, the present study analyzed four VEGF
gene polymorphisms and their impact on the prognosis for
patients with gastric adenocarcinoma.
materials and methods
study population
All the tissues investigated in this study were obtained from consecutive
Korean patients (n = 503) who had undergone a surgical gastrectomy
between January 2000 and December 2001 at Kyungpook National
University Hospital (Daegu, Korea). Written informed consent for gene
expression analyses including SNPs was received from all the patients before
surgery, and the study was approved by the Institutional Research Board at
Kyungpook National University Hospital. The diagnosis and staging of
gastric carcinoma were assessed according to the World Health
Organization classifications [27] and tumor–node–metastasis (TNM)
classifications set out by the American Joint Committee on Cancer [28].
genotyping of VEGF gene polymorphisms
The genomic DNA was extracted from paraffin-embedded tumor bearing
tissue using a Wizard genomic DNA purification kit (Promega, Madison,
WI). The VEGF 2460T > C, 2116G > A, +405G > C, and +936C> T
polymorphisms were then determined using a polymerase chain reaction-
restriction fragment length polymorphism assay. Based on the GenBank
reference sequence (accession no. NT_007592), the PCR primers designed
for 2460T > C, 2116G > A, +405G > C, and +936C > T were 5#-
CCTCTTTAGCCAGAGCCG-GGG-3# (forward) and 5#-TGGCCTTCTCC
CCGCTCCGAC-3# (reverse); 5#-TCCTGCTCCCTCCTCGCCAAATG-3#
(forward) and 5#-GGCGGGGACAGGCGAGCCTC-3# (reverse); 5#-CGAC
GGCTTGGGGAGATTGC-3# (forward) and 5#-GGGCGGTGTCTGTCT
GTCTG-3# (reverse); and 5#-CTCGGTGATTTAGCAGCAAG-3# (forward)
and 5#-CTCGGTGATTTAGCAGCAAG-3# (reverse), respectively. The PCR
reactions were carried out in a 20-ll reaction volume containing 100 ng
genomic DNA, 25 pmol/l each primer, 0.2 mmol/l deoxyuceotide
triphosphates, 10 mmol/l Tris–HCl (pH 8.3), 50 mmol/l KCl, 1.5 mmol/l
MgCl2, and 1 unit of Taq polymerase (Takara Shuzo Co., Otsu, Shiga,
Japan). The PCR cycle conditions consisted of an initial denaturation step
at 94
C for 5 min, followed by 35 cycles of 30 s at 94
C, 30 s at 62
, 30 s at
72
C, and a final elongation at 72
C for 10 min. The PCR products were
digested overnight with the appropriate restriction enzymes (New England
BioLabs, Beverly, MA), which were BsaHI, MnlI, BsmFI, and NlaIII for
the 2460T > C, 2116G > A, +405G > C, and +936C> T polymorphisms,
respectively. The digested PCR products were resolved on a 3% agarose gel
and stained with ethidium bromide for visualization under UV light. For
quality control, the genotyping analysis was done blind as regards the
subjects. The selected PCR-amplified DNA samples (n = 2, for each
genotype) were also examined by DNA sequencing to confirm the
genotyping results.
immunohistochemistry
Formalin fixed, paraffin-embedded serial sections (4 lm) mounted on
3-aminopropyltriethoxysilane-coated slides (Matsunami Glass Ind. Ltd,
Volume 18 | No. 6 | June 2007
original article
Kasamatsu, Japan) were deparaffinized in xylene alcohol and graded
alcohol. Rabbit polyclonal anti-VEGF (Santa Cruz Biotechnology, Inc.,
Santa Cruz, CA) and anti-VEGF-C (Zymed Laboratories, San Francisco,
CA) antibodies were used at concentrations of 1.6 lg/ml and 2.0 lg/ml,
respectively. The immunohistochemistry was carried out using a catalyzed
signal amplification system (Dako, Ely, UK) according to the
manufacturer’s protocol. The sections were treated with 0.3% hydrogen
peroxide (H2O2) in water for 10 min to quench any endogenous peroxidase
activity within the tissue, and the nonspecific binding sites blocked with
a 20% heat-inactivated nonserum protein for 10 min at room temperature.
Next, the sections were incubated for 15 min in the presence of the primary
antibody, then the slides were washed in phosphate buffered saline (PBS)
containing 0.1% Tween 20 (PBS/Tween) for 15 min while changing the
solution three times before the application of the secondary biotinylated
antibody. The slides were incubated with the secondary antibody for 15 min
at room temperature before being washed for 15 min in PBS/Tween that
was changed three times. The sections were then incubated for 15 min with
an avidin–biotinylated horseradish peroxidase complex, and the reaction
visualized with 0.02% 3,3#-diaminobenzidine tetrahydrochloride as
a chromogen in a Tris–HCl buffer, pH 7.6, containing 0.03% H2O2.
Hematoxylin was used to counterstain the nuclei. The staining results for
VEGF and VEGF-C were classified by estimating the percentage of epithelial
cells showing specific immunoreactivity: grade 0 (0%–10% positive cells),
grade I (10%–50% positive cells), and grade II (>50% positive cells). Two
pathologists (HB and GY) blindly examined all the slides.
statistical analysis
The genotypes for each SNP were analyzed as a three-group categoric
variable (referent model), and those were also grouped according to the
dominant and recessive model. The haplotypes and their frequencies were
estimated using the Bayesian algorithm in the phase program [29], which is
available at http://www.stat.washington.edu/stephens/phase.html. The
correlation between the level of VEGF or VEGF-C protein expression and
four VEGF gene polymorphisms was analyzed using chi-square test and
logistic regression analysis. Overall survival (OS) was measured from the
day of surgery until the date of death or last follow-up. Disease-free survival
(DFS) was calculated from the day of surgery until recurrence or death
from any cause. The survival estimates were calculated using the
Kaplan–Meier method. The differences in OS or DFS according to the four
VEGF gene polymorphisms were compared using log-rank tests. Coxs
proportional hazards regression model was used for the multivariate
survival analyses, and the analyses were always adjusted for age (<60 versus
‡60 years), sex (male versus female), stage (0–IV), and adjuvant
chemotherapy (i.v. versus oral versus not received). The hazard ratio (HR)
and 95% confidence interval (CI) were also estimated. The multivariate
analyses were carried out on 492 out of 503 patients, as haplotype data were
unavailable for 11 patients.
A cut-off P value of 0.05 was adopted for all the statistical analyses.
The statistical data were obtained using an SPSS software package
(SPSS 11.5 Inc., Chicago, IL) or SAS Genetic software (SAS Institute,
Cary, NC).
results
patient characteristics and survival analysis
The median age of the patients was 60 (range, 25–83) years,
and 337 (67.0%) patients were male. Curative resections
(gastrectomy with more than D1 lymph node dissection) were
carried out in 479 (97.2%) patients, while the others received
a palliative gastrectomy. The pathologic stages after the surgical
doi:10.1093/annonc/mdm085 | 1031
original article Annals of Oncology

resection were as follows: stage 0 (n = 5, 1.0%), stage IA
(n = 172, 34.2%), stage IB (n = 106, 21.1%), stage II (n = 105,
20.9%), stage IIIA (n = 47, 9.3%), stage IIIB (n = 26, 5.2%),
and stage IV (n = 42, 8.3%). Among the 220 patients
with stage II–IV diseases, 167 (75.9%) patients received
adjuvant chemotherapy with four cycles of 5-fluorouracil
(5-FU) + epirubicin (n = 69) or nimustine (n = 12) followed
by oral 5-FU for 1 year or just oral 5-FU for 1 year (n = 86).
At the time of last analysis (January 2006), 111 patients had
experienced a disease relapse and 103 patients had died as
a result of gastric cancer. The death of 6 patients, however,
was not related to gastric cancer. The estimated 5-year
OS and DFS for all the patients was 78.4 6 1.85% and
77.4 6 1.87%, respectively.
genotype frequency and effects on survival
The four VEGF gene polymorphisms were successfully
amplified in 98%–99% of the cases. The frequencies of each
genotype are shown in Table 1. The 2460T > C and +405G > C
or +936C > T polymorphisms exhibited a strong linkage
disequilibrium (correlation coefficient, R = 0.26; Lewontin’s D’,
D’ = 0.79 or correlation coefficient, R = 0.39; Lewontin’s D’,
D’ = 0.91) in the study population, whereas the linkage with
the 2116G > A polymorphisms was weaker (correlation
coefficient, R = 0.09; Lewontin’s D’, D’ = 0.65). The survival
analysis showed no association between the three VEGF gene
polymorphisms (2460T > C, 2116G > A, and +405G > C) and
the prognosis. For the +936C > T polymorphism, the T/T
genotype, however, exhibited a worse OS compared with the
C/C genotype (HR = 3.23; 95% CI, 1.13–9.25; P = 0.037),
although no significant difference was observed in the
dominant or recessive model (Table 1). When confined to the
patients with stage 0 or I gastric cancer, the OS and DFS for
the patients with the 2460 T/C or C/C genotype (dominant
model) were worse than those for the patients with
the 2460 T/T genotype in the multivariate analysis
(OS: HR = 3.96, 95% CI, 1.08–14.52, P = 0.038;
DFS: HR = 4.87, 95% CI, 1.36–17.43, P = 0.015, Table 2
and Figure 1), meanwhile the other polymorphisms had
no significant effect on survival.
haplotype analysis
The haplotype analyses were conducted to evaluate the
combined effect of the four polymorphisms on gastric cancer
survival. Fourteen haplotypes (TGGC 46.2%; TGGT 7.9%;
TGCC 17.0%; TGCT 0.2%; TAGC 0.7%; TACC 2.0%; CGGC
4.1%; CGGT 0.1%; CGCC 5.6%; CGCT 3.2%; CAGC 0.9%;
CAGT 1.3%; CACC 6.5%; and CACT 4.3%;) were estimated
from the four polymorphisms in 492 (97.8%) out of 503 cases.
Since TGGC, TGGT, TGCC, CGCC, and CACC were the
common haplotypes, the impact of these haplotypes on survival
was analyzed (Table 3). The estimated 5-year OS and DFS rates
for the patients with the CACC haplotype were 73.4 6 5.5%
and 74.6 6 5.5%, respectively, which were significantly inferior
to the rates for the patients with the TGGC haplotype
(OS: HR = 1.72, 95% CI, 1.02–2.89; P = 0.041,
DFS: HR = 1.73, 95% CI, 1.03–2.92; P = 0.038, Table 3
and Figure 2). In a Cox model for OS that included the age,
TNM stage, and adjuvant chemotherapy, the TNM stage
remained a significant prognostic factor (P < 0.001).
immunostaining of VEGF proteins
The VEGF and VEGF-C protein expression was evaluated
by immunohistochemical staining in 374 (74.4%) and
371 (73.8%) cases, respectively. No correlation was detected
with the level of VEGF or VEGF-C protein expression and any
of the four VEGF gene polymorphisms analyzed (data not

Table 1. Overall and disease-free survival according to four VEGF gene polymorphisms in all patients

Genotype No. (%) Overall survival Disease-free survival

HR 95% CI P value HR 95% CI P value
2460T > C polymorphism 500
T/T 259 (51.8) 1 0.107 1 0.115
T/C 224 (44.8) 1.264 0.852–1.876 1.280 0.874–1.875
C/C 17 (3.4) 2.596 1.003–6.719 2.463 0.956–6.342
2116G > A polymorphism 496
G/G 339 (68.3) 1 0.191 1 0.150
G/A 157 (31.7) 1.300 0.878–1.926 1.323 0.904–1.936
A/A 0
+405G > C polymorphism 499
G/G 150 (30.1) 1 0.389 1 0.495
G/C 312 (62.5) 0.749 0.495–1.134 0.790 0.527–1.184
C/C 37 (7.4) 0.790 0.373–1.675 0.769 0.365–1.623
+936C > T polymorphism 498
C/C 337 (67.7) 1 0.037 1 0.047
C/T 152 (30.5) 0.777 0.500–1.209 0.828 0.542–1.266
T/T 9 (1.8) 3.231 1.128–9.252 3.270 1.145–9.342

P values correspond to multivariate Cox model adjusted for age, sex, tumor–node–metastasis stage, and adjuvant chemotherapy.
VEGF, vascular endothelial growth factor; HR, hazard ratio; CI, confidential interval.

1032 | Kim et al. Volume 18 | No. 6 | June 2007

Annals of Oncology original article
Table 2. Overall and disease-free survival according to 2460T > C genotype in patients with early stage gastric cancer (stage 0 or 1)

Genotype No. (%) Overall survival Disease-free survival

HR 95% CI P value HR 95% CI P value
2460T > C polymorphism 282
Referent model
T/T 147 (52.1) 1 0.104 1 0.046
T/C 127 (45.0) 4.102 1.118–15.046 5.029 1.406–17.993
C/C 8 (2.8) Undefined No events Undefined No events
Dominant model for C alleles
T/T 147 (52.1) 1 0.038 1 0.015
T/C–C/C 135 (47.9) 3.957 1.078–14.521 4.870 1.360–17.432

P values correspond to multivariate Cox model adjusted for age, sex, and tumor–node–metastasis stage.
HR, hazard ratio; CI, confidential interval.

shown). Also, no correlation was observed with any other A

clinicopathologic variables. Finally, no association between
the grade of immunostaining and the OS or DFS was
observed (Table 4).

discussion
We have investigated the prognostic impact of four VEGF gene
polymorphisms in quite a large population of patients with
surgically resected gastric adenocarcinoma. The current study
demonstrated that CAGG haplotype or 2460 T/C–C/C
genotype was a poor prognostic factor in these patients.
The role of genetic polymorphisms, which is an important T/T
T/C-C/C
determinant of endogenous causes of cancer, in relation to the
P=0.038
risk for gastric cancer has attracted increasing interest,
probably because of advances in DNA analysis technologies and
human genome knowledge. Most previous studies, however,
have only addressed the effect of genetic variants of metabolic
enzymes and inflammation mediators [30]. Since VEGF or B
its family plays a critical role in tumor-related angiogenesis,
the association of VEGF gene polymorphisms with the risk or
prognosis of several solid tumors, such as melanoma [18],
lung [19], prostate [20], and breast cancer [21, 23, 24], has
already been demonstrated.
The genotype frequencies of 2460T > C, +405G > C
or +936C > T in the present study were similar to those
previously reported for the Korean [19], Chinese [23],
and European populations [16, 21, 24], whereas the
frequency of the 2116 G/G genotype was higher than
that for European patients (68.3% versus 42.1%) [24]. T/T
T/C-C/C
The 2460T > C and +405G > C polymorphisms exhibited
a strong linkage disequilibrium in the present study population, P=0.015

while the linkage between the 2460T > C and 2116G > A
polymorphisms was weak, which is consistent with findings
from an earlier study [15, 19, 23].
Figure 1. Overall survival (A) and disease-free survival (B) curves
In the present study, the CACC haplotype was associated
according to 2460T > C polymorphism in patients with stage 0 or I gastric
with a significantly worse survival when compared with the
cancer. P values correspond to dominant model in multivariate Cox model
TGGC haplotype in patients with surgically resected gastric
adjusted for age, sex, and tumor–node–metastasis stage.
cancer (OS: HR = 1.72, DFS: HR = 1.73), and the 2460 T/C
or C/C genotype was a poor prognostic factor in patients
with stage 0 or I gastric cancer (OS: HR = 3.957, (2460T > C, +405G > C, and +936C> T) on survival
DFS: HR = 4.870). In a recent study by Lu et al. [23] that in 1193 Chinese patients with breast cancer, the survival
also evaluated the effects of three VEGF gene polymorphisms was lower among the patients with the 2460 C/C

Volume 18 | No. 6 | June 2007 doi:10.1093/annonc/mdm085 | 1033

original article Annals of Oncology

Table 3. Overall and disease-free survival according to common haplotypes of four VEGF gene polymorphisms in all patients

Haplotypes (2460T > Haplotype Overall survival Disease-free survival

C/2116G > A/+405G > frequency (%) HR 95% CI P value HR 95% CI P value
C/+936C > T)
TGGC 46.2 1 1
TGGT 7.9 1.160 0.667–2.017 0.600 1.240 0.723–2.126 0.434
TGCC 17.0 0.995 0.659–1.502 0.979 1.047 0.701–1.563 0.824
CGCC 5.6 1.251 0.681–2.299 0.471 1.129 0.600–2.124 0.707
CACC 6.5 1.719 1.022–2.893 0.041 1.733 1.031–2.915 0.038

P values correspond to multivariate Cox model adjusted for age, sex, tumor–node–metastasis stage, and adjuvant chemotherapy.
VEGF, vascular endothelial growth factor; HR, hazard ratio; CI, confidential interval.

and +405 G/G genotypes and higher among the patients with A
the 2460T/+405C/+936C haplotype. One possible explanation
for these results is that the DNA sequence variations in the
VEGF gene may alter VEGF production and/or activity, thereby
causing interindividual differences in the lymphangiogenesis
and lymphatic tumor spread. Given the homogenous ethnic
background of Korean patients, any potential confounding
effect due to ethnicity is likely to be small in the present study.
The correlation between an elevated preoperative serum VEGF
concentration and the poor survival in patients with gastric
cancer was already demonstrated in previous study [31]. A few
studies, however, have reported that VEGF gene TGGC
polymorphisms are associated with VEGF production. CACC
Nonetheless, the results are inconsistent. Awata et al. [32] P=0.041
reported that individuals with the +405 C/C genotype had
a higher fasting serum VEGF level than those with other
genotypes, and that they carried an increased risk of
diabetic retinopathy. Meanwhile, Watson et al. [15]
documented that the +405G allele is associated B
with higher lipopolysaccharide-stimulated VEGF production
by peripheral blood mononuclear cells than the +405C
allele. In a recent in vitro study [17], carrying a haplotype
containing the 2460C/+405G polymorphisms was found to
significantly increase basal VEGF promoter activity and
phorbol ester-induced responsiveness compared with the
presence of a haplotype containing the 2460T/+405C
polymorphisms.
In the present study, any of the four VEGF gene
polymorphisms was not correlated with the level of VEGF
TGGC
or VEGF-C protein expression. Koukourakis et al. [33] CACC
reported that 22578 C/C, 2634 G/G, and 21154 A/A and G/A
P=0.038
genotype (from translation start site) in the VEGF gene were
linked with low VEGF protein expression in non-small cell
lung cancer (NSCLC). They indicate that inherited variations
in VEGF sequence, that also characterize the tumor genome, Figure 2. Overall survival (A) and disease-free survival (B) curves
are determinants of the molecular VEGF phenotype in according to haplotypes in all patients with gastric cancer. P values
NSCLC and, consequently, of the intratumoral vascular correspond to multivariate Cox model adjusted for age, sex,
density. Sample size, however, was small (36 cases) and tumor–node–metastasis stage, and adjuvant chemotherapy.
VEGF 2460T > C, +405G > C, and +936C> T polymorphisms
were not evaluated in their study. Accordingly, further
studies are needed to elucidate the relationship between the into clinical trials for different tumor types, and the results are
VEGF gene polymorphisms and serum VEGF level or VEGF very encouraging [34]. Thus, if the prediction of individual
protein expression. angiogenic potential on the basis of VEGF genotypes is possible,
Recently, several VEGF inhibitors, including bevacizumab, the efficacy of antiangiogenic treatment could be further
a monoclonal antibody against VEGF, have actively progressed enhanced.

1034 | Kim et al. Volume 18 | No. 6 | June 2007

Annals of Oncology original article
Table 4. Survival analysis according to the level of VEGF and VEGF-C protein expression

No. % Overall survival Disease-free survival

HR (95% CI) P value HR (95% CI) P value
Tumor cells expressing VEGF 374
Grade 0 (£10%) 37 9.9 1 0.234 1 0.191
Grade I/II (>10%) 337 91.1 1.747 (0.696–4.384) 1.844 (0.737–4.617)
Tumor cells expressing VEGF-C 371
Grade 0 (£10%) 93 25.1 1 0.076 1 0.133
Grade I/II (>10%) 278 74.9 1.562 (0.954–2.558) 1.429 (0.897–2.277)

P values correspond to multivariate Cox model adjusted for age, sex, tumor–node–metastasis stage, and adjuvant chemotherapy.
VEGF, vascular endothelial growth factor; VEGF-C, vascular endothelial growth factor C; HR, hazard ratio; CI, confidential interval.

In conclusion, VEGF gene polymorphisms were found to
be an independent prognostic marker for Korean patients with
surgically resected gastric adenocarcinoma. Consequently, in
addition to the pathologic stage, the analysis of VEGF gene
polymorphisms may help identify patient subgroups at high
risk for a poor disease outcome, thereby helping to refine
therapeutic decisions in gastric cancer.
acknowledgements
This work was supported by BioMedical Research Institute
grant, Kyungpook National University Hospital (2005). This
work was presented at the 42nd Annual Meeting of the
American Society of Clinical Oncology, Atlanta, GA,
June 2–6, 2006.
references
1. Kim JP, Hur YS, Choe KJ et al. Clinical analysis of 1136 early gastric cancers.
Cancer Res Treat 1993; 25: 808–818.
2. Parkin DM. Global cancer statistics in the year 2000. Lancet Oncol 2001;
2: 533–543.
3. Yancopoule GD, Davis S, Gale NW et al. Vascular-specific growth factors and
blood vessel formation. Nature 2000; 407: 242–248.
4. Carmeliet P, Jain RK. Angiogenesis in cancer and other disease. Nature 2000;
407: 249–257.
5. Toi M, Matsumoto T, Bando H. Vascular endothelial growth factor: its prognostic,
predictive, and therapeutic implications. Lancet Oncol 2001; 2: 667–673.
6. Masuya D, Huang C-H, Liu D et al. The intratumoral expression of vascular
endothelial growth factor and interleukin-8 asscociated with angiogenesis in
nonsmall cell lung carcinoma patients. Cancer 2001; 92: 2628–2638.
7. Fontanini G, Faviana P, Lucchi M et al. A high vascular count and overexpression
of vascular growth factor are associated with unfavorable prognosis in operated
small cell lung cancer. Br J Cancer 2002; 86: 558–563.
8. Nishida N, Yano H, Komai K et al. Vascular endothelial growth factor C and
vascular endothelial growth factor receptor 2 are related closely to the prognosis
of patients with ovarian carcinoma. Cancer 2004; 101: 1364–1374.
9. Stacker SA, Achen MG, Jussila L et al. Lymphangiogenesis and cancer
metastasis. Nat Rev Cancer 2002; 2: 573–583.
10. Ichikura T, Tomimatsu S, Ohkura E et al. Prognostic significance of the
expression of vascular endothelial growth factor (VEGF) and VEGF-C in gastric
carcinoma. J Surg Oncol 2001; 78: 132–137.
11. Yonemura Y, Endo Y, Fujita H et al. Role of vascular endothelial growth factor
C expression in the development of lymph node metastasis in gastric cancer.
Clin Cancer Res 1999; 5: 1823–1829.
12. Juttner S, Wissmann C, Jons T et al. Vascular endothelial growth factor-D and
its receptor VEGFR-3: two novel independent prognostic markers in gastric
adenocarcinoma. J Clin Oncol 2006; 24: 228–240.
13. Mattei MG, Borg JP, Rosnet O et al. Assignment of vascular endothelial
growth factor (VEGF) and placenta growth factor (PIGF) genes to human
chromosome 6p12-p21 and 14q24-q31 regions, respectively. Genomics 1996;
31: 168–169.
14. Tischer E, Mitchell R, Hartman T et al. The human gene for vascular endothelial
growth factor: multiple protein forms are encoded through alternative exon
splicing. J Biol Chem 1991; 266: 11947–11954.
15. Watson CJ, Webb NJA, Bottomley MJ, Brenchley PEC. Identification of
polymorphisms within the vascular endothelial growth factor (VEGF) gene:
correlation with variation in VEGF protein production. Cytokine 2000;
12: 1230–1235.
16. Renner W, Kotschan S, Hoffman C et al. A common 936C/T mutation in the gene
for vascular endothelial growth factor is associated with vascular endothelial
factor plasma levels. J Vasc Res 2000; 37: 443–448.
17. Stevens A, Soden J, Brenchley PE et al. Haplotype analysis of the polymorphic
human vascular endothelial growth factor gene promoter. Cancer Res 2003;
63: 812–816.
18. Howell WM, Bateman AC, Turner SJ et al. Influence of vascular endothelial
growth factor single nucleotide polymorphisms on tumor development in
cutaneous malignant melanoma. Genes Immun 2002; 3: 229–232.
19. Lee SJ, Lee SY, Jeon HS et al. Vascular endothelial growth factor gene
polymorphisms and risk of primary lung cancer. Cancer Epidemiol Biomarkers
Prev 2005; 14: 571–575.
20. Lin CC, Wu HC, Tsai FJ et al. Vascular endothelial growth factor gene
2460C/T polymorphism is a biomarker for prostate cancer. Urology 2003;
62: 374–377.
21. Krippl P, Langsenlehner U, Renner W et al. A common 936 C/T gene
polymorphism of vascular endothelial growth factor is associated with decreased
breast cancer. Int J Cancer 2003; 106: 468–471.
22. Ruzzo A, Graziano F, Kawakami K et al. Pharmacogenetic profiling and clinical
outcome of patients with advanced gastric cancer treated with with palliative
chemotherapy. J Clin Oncol 2006; 24: 1883–1891.
23. Lu H, Shu X, Cui Y et al. Association of genetic polymorphisms in the VEGF gene
with breast cancer survival. Cancer Res 2005; 65: 5015–5019.
24. Jin Q, Hemminki K, Enquist K et al. Vascular endothelial growth factor
polymorphisms in relation to breast cancer development and prognosis.
Clin Cancer Res 2005; 11: 3647–3653.
25. Dotor E, Cuatrecases M, Martinez-Iniesta M et al. Tumor thymidylate
synthase 1496del6 genotype as a prognostic factor in colorectal cancer
patients receiving flourouracil-based adjuvant treatment. J Clin Oncol 2006
24: 1603–1611.
26. Li D, Frazier M, Evans DB et al. Single nucleotide polymorphisms of RecQ1,
RAD54L, and ATM genes are associated with reduced survival of pancreatic
cancer. J Clin Oncol 2006; 24: 1720–1728.
27. Hamilton SR, Aaltonen LA. WHO Classification. Tumours of the Digestive System:
Pathology & Genetics. Lyon, France IARC Press 2000.
28. Greene FL, Page DL, Fleming ID et al. The AJCC cancer staging manual,
6th edition. New York, NY: Springer-Verlag 2002.
29. Stephens M, Smith MJ, Donnelly P. A new statistical method for haplotype
reconstruction from population data. Am J Hum Genet 2001; 68: 978–989.

Volume 18 | No. 6 | June 2007 doi:10.1093/annonc/mdm085 | 1035

original article
30. Gonzalez CA, Sala N, Capella G. Genetic susceptibility and gastric cancer risk.
Int J Cancer 2002; 100: 249–260.
31. Karayiannakis AJ, Syrigos KN, Polychronidis A et al. Circulating VEGF levels in
the serum of gastric cancer patients. Ann Surg 2002; 236: 37–42.
32. Awata T, Inoue K, Kurihara S et al. A common polymorphism in the
5#-untranslated region of the VEGF gene is associated with diabetic retinopathy
in type 2 diabetes. Diabetes 2002; 51: 1630–1635.
1036 | Kim et al.
Annals of Oncology
33. Koukourakis MI, Papazoglou D, Giatromanolaki A et al. VEGF gene
sequence variation defines VEGF gene expression status and
angiogenic activity in non-small cell lung cancer. Lung Cancer 2004; 46:
293–298.
34. Hurwitz H, Fehrenbacher L, Novotny W et al. Bevacizumab plus irinotecan,
fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 2004;
350: 2335–2342.
Volume 18 | No. 6 | June 2007