This study comprised 104 patients who underwent surgical resection for gastric adenocarcinoma at Hanyang University Guri Hospital between April 2000 and November 2003. Of those, 84 patients had advanced gastric cancers and 20 patients had early gastric cancers. Well-documented clinical data were collected from all patients. Information concerning the date of initial diagnosis, clinical characteristics, relapse, and death were retrospectively obtained. In addition, adjacent non-neoplastic stomach tissue samples as confirmed by Hematoxylin and Eosin staining were used as controls. This study was approved by the institutional review board of Hanyang University Guri Hospital.

The avidin-biotin complex (ABC) method was used for immunostaining. Formalin-fixed, paraffin-embedded tissue blocks were sectioned at a 4-µm thickness. The tissue sections were deparaffinized by three, 10-min incubations in xylene and then rehydrated in serial graded alcohol. For antigen retrieval, the sections were heated in a microwave oven for 10 min in 10 mM/L sodium citrate buffer (pH 6). Endogenous peroxidase activity was eliminated by preincubation in 3% hydrogen peroxide and 10% methanol for 15 min followed by three washes in phosphate-buffered saline. All slides were pre-incubated at 37℃ for 20 min with two drops of normal blocking solution (goat serum, 100 µL/slide). The slides were then incubated with either a goat polyclonal anti-VEGF-D antibody (R&D Systems, Minneapolis, MN) at a 1:100 dilution overnight at 4℃ or a rabbit polyclonal anti-VEGFR-3 antibody (Zymed Laboratories, San Francisco, CA, U.S.A.) at a 1:200 dilution for 2 hr at room temperature. Biotinylated secondary antibody was added to each slide and incubated for 30 min at 37℃. The slides were then treated with the avidin-biotinylated peroxidase complex (Immunotech, Cedex, France) for an additional 30 min at room temperature. 3, 3'-diaminobenzidine tetrahydrochloride (Immunotech, Cedex, France) was used for color development. Finally, the sections were counterstained with hematoxylin.

All slides were coded and evaluated by two experienced pathologists without knowledge of patient identity or clinical status. Each experiment was performed twice independently. In the discrepant cases, two pathologists reviewed the cases together and reached a consensus. The VEGF-D staining intensity was scored as 0 (negative), 1 (weak), 2 (medium), and 3 (strong). The extent of staining was scored according to the percentage of areas with positive VEGF-D staining as follows: 0 (0%), 1 (1 to 25%), 2 (26 to 50%), 3 (51 to 75%), and 4 (76 to 100%). The final staining score was derived from the sum of the intensity score and the extent score. Tumors with a final staining score of ≥5 were considered as positive for VEGF-D expression.

The VEGFR-3-positive vessel density was assessed according to the method described by Weidner et al. (19). Microvessel counting was performed twice. Each slide was first scanned at 100× magnification to determine three "hot spots" defined as areas with the maximum number of VEGFR-3-positive vessels. The VEGFR-3-positive vessel density was determined by counting all the immunostained vessels at a 200× magnification and the mean number of positive vessels was calculated in the three selected areas for each case.

The Pearson chi-square (χ²) test was performed to determine the correlation between VEGF-D expression and various clinicopathological factors. The Mann-Whitney U test and Kruskal-Wallis test were used to examine the association of VEGFR-3-positive vessel density. Survival curves were calculated using the Kaplan-Meier method and compared with other prognostic variables using the log-rank test. A stepwise Cox's regression analysis was performed to identify prognostic factors for survival. In all tests, p<0.05 was considered statistically significant. All statistical analyses were performed using SPSS 10® statistical software.

Our study included 71 male (68.3%) and 33 female (31.7 %) patients, with a median age of 59 yr (range, 25-79 yr). All tumors were staged using the AJCC classification. Forty-three patients (41.4%) were classified as stage I, 11 patients (10.6%) as stage II, 20 patients (19.2%) as stage III, and 30 patients (28.8%) as stage IV. R0 resection was done in 95 patients, R1 resection in 4, and R2 resection in 5. Forty-one patients received intravenous systemic chemotherapy after surgery (17 patients: cisplatin+5-fluorouracil (5-FU), 16 patients: heptaplatin +5-FU, and 8 patients: paclitaxel+cisplatin+5-FU) and 56 patients received 5-fluorouracil orally. The median follow-up period was 35.9 months (range: 2.1-70.5 months). Thirty-five patients had relapsed by the time of last follow-up and thirty-five patients died. The most common cause of death was disease progression (30 patients), while other causes of death included respiratory failure (1 patient), bowel infarction (1 patient), liver cirrhosis (1 patient), malnutrition (1 patient), and septic shock (1 patient).

A granular pattern of VEGF-D staining was observed within the cytoplasm of malignant gastric epithelium (Fig. 1). VEGF-D expression was observed in 62.5% of gastric cancers and in 9.6% of non-neoplastic gastric tissues (p<0.001). VEGF-D expression was significantly associated with lymph node metastasis (p<0.05), increased serum carcino embryonic antigen (CEA) levels (p<0.05), and the non-signet ring cell type (p<0.001, Table 1). However, we found no significant differences in other clinicopathological parameters such as age, sex, primary tumor size, grade, lymphatic and venous invasion, depth of tumor invasion (T stage), distant metastasis, intravenous systemic chemotherapy, curative respectability (R0/R1, R2), and stage between the VEGF-D-positive and -negative groups.

We observed VEGFR-3 expression on endothelial cells. Most of the VEGFR-3-stained vessels were thin-walled and did not contain red blood cells (Fig. 1). The VEGFR-3 positive vessel density was significantly greater in gastric carcinomas (median, 49.5; range, 0-205) than in non-neoplastic tissues (26.0; 0-98). Furthermore, the VEGFR-3-positive vessel densities were 60.0 (10-205) and 36 (0-100) in the VEGF-D positive and negative groups, respectively (p<0.001, Fig. 2). The VEGFR-3 positive vessel density was significantly correlated with primary tumor size (p<0.05), lymphatic invasion (p<0.05), and lymph node metastasis (p<0.05, Table 1). However, there was no significant difference in other clinicopathological parameters such as age, sex, pathologic type, grade, venous invasion, depth of tumor invasion, and distant metastasis.

To evaluate the correlation between VEGFR-3-positive vessel density and patient survival, the patients were divided into two groups: low (<50) and high VEGFR-3-positive vessel density (≥50). The cutoff value for each group was determined by the median VEGFR-3-positive vessel density. Univariate analysis showed that tumor size (p<0.001), lymphatic invasion (p<0.05), venous invasion (p<0.05), T (p< 0.001), N (p<0.001), M (p<0.001), and curative resectability (p<0.001) were significant factors for overall survival. However, VEGF-D expression (p>0.05) and high VEGFR-3-positive vessel density (p>0.05) did not influence overall survival. Patients with VEGF-D expression had a significantly shorter relapse-free survival time compared to patients who were VEGF-D-negative (5 yr relapse-free survival rate 55.8% vs. 67.6%, p<0.05, Fig. 3). Furthermore, the high VEGFR-3-positive vessel density group showed a significantly worse relapse-free survival when compared to the low positive vessel density group (5 yr relapse-free survival rate 48.9% vs. 74.9%, p=0.037, Fig. 3). In addition, increased serum CEA levels (p<0.05), tumor size (p<0.001), lymphatic invasion (p<0.001), venous invasion (p<0.05), T-category (p<0.001), N-category (p<0.001), M-category (p<0.001), and curative resectability (p<0.001) were significant prognostic factors for relapse-free survival.

In the multivariate analysis, TNM stage and curative resectability were the independent prognostic factors for disease-free survival and overall survival. The expression of VEGF-D and high VEGFR-3-positive vessel density did not influence survival.