I. Hyodo et al., Clinical significance of plasma vascular endothelial growth factor in gastrointestinal cancer, EUR J CANC, 34(13), 1998, pp. 2041-2045
Circulating vascular endothelial growth factor (VEGF) was measured in gastr
ic and colorectal cancer patients using an enzyme-linked immunosorbent assa
y (ELISA). Firstly, serum and plasma samples were collected from 20 normal
controls to compare the values of VEGF and to determine which specimen type
was most suitable for detecting circulating VEGF. Seventeen of 20 normal c
ontrols had plasma VEGF levels under the limit of detection (15 pg/ml) and
the levels of the remaining three controls were 21, 22 and 38 pg/ml. In con
trast, all serum samples indicated high levels of VEGF (mean 238 pg/ml), ra
nging from 44 to 450 pg/ml. In a time-course test of two normal controls se
rum VEGF values increased markedly between 30 and 60 min and remained high,
whilst plasma VEGF values were low up to 480 min. Thus, plasma samples are
more suitable for the measurement of circulating VEGF. Next, plasma VEGF l
evels were examined in 44 patients with gastric cancer (8 early, 7 advanced
without remote metastasis and 29 metastatic), 13 with colorectal adenoma (
2 with focal cancer) and 26 with colorectal carcinoma (8 advanced without m
etastasis and 18 metastatic) before treatment. An extremely high plasma con
centration of VEGF was seen in some cancer patients with metastasis. To dis
criminate these patients, a cut-off level was determined, considering both
the distribution of the sample concentration and the upper limit of 95% con
fidence area of VEGF in the cancer patients without metastasis. The cut-off
value was 108 pg/ml and most cancer patients without metastases had VEGF l
evels below the cut-off value. In 11 of 29 metastatic gastric cancer patien
ts (38%) and 9 of 18 metastatic colorectal cancer patients (50%), plasma VE
GF levels were higher than the cut-off value. Survival was also analysed in
the patients with metastasis. It was significantly longer in the patients
with low VEGF levels (below the cut-off) than in those with high VEGF level
s (logrank test, P= 0.042). 34 patients with metastasis (19 gastric cancer
and 15 colorectal cancer) were treated with systemic chemotherapy, and thei
r pretreatment levels of plasma VEGF and conventional tumour markers (CEA a
nd CA19-9) were evaluated in relation to response. The response to chemothe
rapy was significantly higher in patients with low VEGF levels (less than o
r equal to 108 pg/ml) than in those with high VEGF levels (P= 0.047). Such
a difference was not observed with CEA/CA19-9. In conclusion, plasma VEGF i
s a useful marker for tumour metastasis and patient survival, and a possibl
e predictive factor for the response of patients with gastrointestinal canc
er to chemotherapy. (C) 1998 Elsevier Science Ltd. All rights reserved.