DOI 10.1245/s10434-010-1033-0
ABSTRACT
Background. Pancreatic cancer is one of the most deadly
cancers, and serum carbohydrate antigen 19-9 (CA19-9)
level has been reported to be a useful prognostic marker in
pancreatic cancer. The purpose of this study was to
determine which prognostic factor (preoperative or postoperative serum CA19-9 level) is more useful.
Methods. Pre- and postoperative serum CA19-9 levels
were measured in 109 patients who underwent surgical
resection for pancreatic cancer between 1998 and 2009,
and their relationships to clinicopathological factors and
overall survival were analyzed with univariate and multivariate methods.
Results. In univariate analysis, tumor location (P =
0.019), postoperative adjuvant chemotherapy (P \ 0.001),
residual tumor factor status (P \ 0.001), UICC pT stage
(P = 0.004), lymph node metastasis (P = 0.015), and
UICC final stage (P = 0.015) were significantly associated
with overall survival. Differences in overall survival were
significant between groups divided on the basis of four
postoperative CA19-9 cutoff values (37, 100, 200, and 500
U/ml) but not significant between groups divided on the
basis of the same four preoperative CA19-9 cutoff values.
Pre- to postoperative increase in CA19-9 level also was
significantly associated with poor prognosis. In multivariate
2322
N. Kondo et al.
Patient Population
From January 1998 to December 2009, a total 115
patients with invasive ductal carcinoma of the pancreas
underwent surgical resection (R0 or R1 resection) at the
Department of Surgery, Hiroshima University Hospital,
Hiroshima, Japan. A diagnosis of pancreatic adenocarcinoma was confirmed histologically in all cases. Other
patients with histological variants, such as mucinous cystic
adenocarcinoma, intraductal papillary adenocarcinoma,
acinar cell carcinoma, and endocrine carcinoma were
excluded.20 Preoperative workup included ultrasonography
(US), computed tomography (CT), endoscopic retrograde
pancreatography, and endoscopic ultrasonography to
evaluate primary and metastatic tumor sites. Patients with
distant metastasis and peritoneal dissemination were
excluded from this analysis even if they were treated by
resection; however, patients with positive paraaortic lymph
nodes only, detected by postoperative histological examination, were included.
Serum CA19-9 level was measured using a CA19-9
radioimmunoassay kit. The manufacturers recommended
value of 37 U/ml was used as the upper limit of normal. In
this study, all patients with obstructive jaundice were
treated with percutaneous transhepatic biliary drainage or
endoscopic retrograde biliary drainage. Preoperative serum
CA19-9 level was measured just before surgery to avoid
the effects of obstructive jaundice, and postoperative
CA19-9 level was measured between 2 and 6 weeks after
surgery. A normal level of serum CA19-9 is defined as \37
U/ml, which was estimated based on the standard deviation
in a normal population. However, individuals with a Lea-bgenotype (lacking the Lewis antigen glycosyl transferase)
are unable to synthesize CA19-9 (approximately 57% of
the general population).21,22 Of 115 patients, 6 patients
(5%) had preoperative and postoperative CA19-9 values \2 U/ml. They were judged to be nonsecretors of
CA19-9 and excluded from this study. Therefore, 109
patients with invasive ductal carcinoma of pancreas who
underwent surgical resection were enrolled in this study.
The clinicopathological characteristics and prognosis were
available for all these patients.
Postoperative adjuvant chemotherapy was administered beginning in 2002 and given to 81 (74%) patients.
Surgical Procedure
All surgical resections included pancreatoduodenectomy (PD), pylorus-preserving pancreatoduodenectomy
(PPPD), distal pancreatectomy (DP), and total pancreatectomy (TP). Patients with carcinoma in the pancreatic
head usually underwent PPPD. However, if the tumor was
close to duodenal bulb area in the superior pancreatic
head, PD with antrectomy was performed. All patients
with carcinoma in the pancreatic body and tail underwent
distal pancreatectomy with splenectomy. All patients
underwent regional lymph node dissection and paraaortic
lymph node dissection. However, to prevent severe postsurgical diarrhea, the nerve plexus around the superior
mesenteric artery was not dissected in any patient. Partial
resection of the portal vein was performed if the surgeon
observed tumor invasion of the portal vein. Intraoperative
pathologic assessment of proximal or distal pancreatic
margins was performed using frozen-tissue sections. If the
pancreatic margin was positive for cancerous cells, further
resection of the pancreas was performed. TP was performed only in cases where negative margins could only
be achieved with TP in preoperative or intraoperative
diagnosis.
Pathological Investigations
Resected specimens were stained with hematoxylin and
eosin and examined histologically, and each tumor was
classified as well differentiated, moderately differentiated,
or poorly differentiated adenocarcinoma according to the
predominant pattern of histological differentiation. Anterior serosal invasion, retropancreatic tissue invasion,
splenic or portal vein invasion, splenic artery invasion,
lymph node metastasis, and extrapancreatic nerve plexus
invasion were all evaluated histologically. Residual tumor
(R factor) was considered R1 if adenocarcinoma infiltrated
the proximal or distal pancreatic transection line or was in
the dissected peripancreatic soft-tissue margins. All
patients with R2 were excluded from this study. Tumor
stage, lymph node metastasis, and final stage were classified based on the 6th edition of the International Union
2323
RESULTS
Demographic Data and Pathological Assessment
Survival
All patients were followed regularly in outpatient clinics
with measurement of CA19-9 every 12 months, and by
performing US of the abdomen and contrast-enhanced CT
scans of the chest, abdomen, and pelvis every 36 months.
Recurrence was defined as radiological evidence of intraabdominal or abdominal soft tissue around the surgical site,
or of distant metastasis. For nonsurvivors, survival time
after surgery and the cause of death were recorded. For
survivors, postsurgical time and status of recurrence were
recorded. Survival analyses on four clinical factors (age,
gender, tumor location, and use of adjuvant chemotherapy)
and five pathological factors (R factor, pathological differentiation, UICC pT stage, lymph node metastasis, and
UICC stage) were performed with univariate and multivariate method.
Statistical Analysis
The patients were divided into two groups on the basis
of clinical and pathological factors. The between-group
difference in median perioperative CA19-9 level was
evaluated by using the Wilcoxon two-sample test. Overall
survival was analyzed by the Kaplan-Meier method, and
significance was determined by the log-rank (Mantel-Cox)
test. The Kaplan-Meier method and log-rank tests were
used to determine difference in overall survival between
each of two groups divided on the basis of clinical and
pathological factors, and four pre- and postoperative
CA19-9 cutoff values (37 U/ml, 100 U/ml, 200 U/ml, and
500 U/ml). These four kinds of cutoff value were determined based on previous reports and quartiles of
preoperative CA19-9 levels (lower: 35 U/ml, median:
122U/ml, and higher: 510 U/ml) and postoperative CA19-9
levels (lower: 9 U/ml, median: 22 U/ml, and higher: 73 U/
ml) in our data.16,17,24 Similarly, we analyzed the difference in overall survival between two groups divided on the
basis of percent decrease in pre- to postoperative CA19-9
levels (increase or decrease, 25%, 50%, and 75%). Fourteen patients with pre- to postoperative increase in CA19-9
level were not included in the \25%, \50%, and \75%
groups.
Factors found significant by univariate statistical analysis were entered into a multivariate proportional hazards
regression model (Cox regression) to determine independent factors predictive for overall survival. All statistical
analysis was performed using JMP statistical software
version 5.1 (SAS Institute, Cary, NC). P \ 0.05 was considered statistically significant.
2324
N. Kondo et al.
70
Range
4387
Gender
Male
56 (51)
Female
53 (49)
Tumor location
Head
78 (72)
Body or tail
31 (28)
Type of resection
PD
4 (3)
PPPD
74 (68)
DP
29 (27)
TP
2 (2)
Adjuvant chemotherapy
Yes
81 (74)
No
28 (26)
77 (71)
R1
32 (29)
Pathological differentiation
Well
Moderately
68 (62)
33 (31)
Poorly
8 (7)
UICC pT stage
T1
9 (8)
T2
9 (8)
T3
91 (84)
69 (63)
No
40 (37)
5 (5)
IB
6 (5)
IIA
26 (24)
IIB
58 (53)
IV
14 (13)
DISCUSSION
Tumor-associated CA19-9 antigen has become the most
important tumor marker for pancreatic cancer. Many
reports have attested to its clinical usefulness in diagnosis,
assessment of resectability, and monitoring of pancreatic
cancer progression and prognosis.2529 In those reports, the
cutoff values of CA19-9 as prognostic indicators were
variable for both preoperative measurements (undetectable,
37 U/ml, 50 U/ml, 370 U/ml, and 2000 U/ml) and postoperative (40 U/ml, 70 U/ml, 90 U/ml, 180 U/ml, 200
U/ml).1519,24,3032 Thus, the preferred prognostic marker
(preoperative or postoperative CA19-9) has not been clear
and the optimal cutoff value has remained controversial.
Multivariate analyses by Ferrone et al. and Montgomery
et al. of both preoperative and postoperative CA19-9 found
that postoperative CA19-9 was a better prognostic marker,
although univariate analysis determined that preoperative
CA19-9 also was a significant prognostic marker.17,33
Postoperative increases in the CA19-9 value and postoperative CA19-9 value [200 U/ml were independent
prognostic factors according to Ferrone et al.17 Postoperative CA19-9 value \180 U/ml at 3 months was an
2325
No. of patients
Preoperative CA19-9
Postoperative CA19-9
Median (U/ml)
Median (U/ml)
P value
23
0.875
P value
Clinical factors
Age (yr)
\70
52
84
C70
57
163
Male
56
100
Female
53
164
Head
78
158
Body or tail
31
97
80
153
29
102
0.682
22
Gender
0.473
24
0.284
20
Tumor location
Type of resection
PD/PPPD/TP
DP
0.650
22
0.979
25
0.870
21
0.666
25
Adjuvant chemotherapy
Yes
81
156
No
28
87
0.542
19
0.351
26
Pathological factors
Residual tumor (R factor)
R0
77
118
R1
32
203
0.194
19
0.041
35
Pathological differentiation
Well
68
113
Moderate/Poorly
41
193
0.691
21
0.584
27
UICC pT stage
T1/T2
PD pancreatoduodenectomy,
PPPD pylorus-preserving
pancreatoduodenectomy,
DP distal pancreatectomy,
TP total pancreatectomy
18
58
T3
91
Lymph node metastasis
164
Yes
69
181
No
40
97
11
96
II/IV
98
160
0.022
20
0.209
24
0.066
27
0.006
14
UICC stage
0.445
20
0.450
23
useful prognostic factor in patients with resectable pancreatic cancer. Moreover, preoperative CA19-9 level was
significantly higher only in patients with UICC pT3.
Generally, CA19-9 has been reported to correlate with
tumor burden,37,38 and higher preoperative CA19-9 may
reflect heavy or extensive tumor burden. Although it seems
to reflect a heavy tumor burden, preoperative CA19-9 level
may not always reflect residual tumor, lymph node
metastasis, and advanced UICC stage. Therefore, in some
cases, high preoperative CA19-9 may reflect the extension
of the tumor macro- or microscopically beyond the margins
of the surgically resected specimen, whereas in other cases,
it may reflect a heavy tumor burden within the limits of
surgical resection. Therefore, we believe that curative
operation should be tried even if the preoperative CA19-9
2326
N. Kondo et al.
P
value
Overall
Survival
1.0
0.8
p = 0.0003
Clinical factors
0.6
Age (yr)
\70
52
44
C70
57
31
Male
56
25
Female
53
50
78
29
0.410
0.4
Gender
0.083
0.2
Tumor location
Head
Body or tail
31
58
Adjuvant chemotherapy
Yes
81
51
No
28
\0.001
Pathological factors
Residual tumor (R factor)
R0
77
46
R1
32
14
\0.001
Pathological differentiation
Well
68
46
Moderately/
Poorly
41
27
T1/T2
18
76
T3
91
24
0.076
UICC pT stage
0.004
0.019
69
40
24
51
0.015
11
79
0.015
II/IV
98
27
Number
at Risk
68
41
13
2
10
0
6
0
Years
44
22
30
4
TABLE 4 Univariate analysis of overall survival in two groups divided on the basis of perioperative CA19-9 cutoff value for 109 patients
treated by surgical resection for pancreatic cancer
CA19-9 cutoff
value (U/ml)
Preoperative CA19-9
No. of
patients
3-Year
survival rate (%)
P value
\37
32
57
0.119
C37
\100
77
49
30
48
C100
60
27
\200
66
45
C200
43
22
\500
81
42
C500
28
13
Postoperative CA19-9
3-Year
survival
rate (%)
P value
68
49
0.0003
0.077
41
89
10
43
0.0019
20
0.194
99
38
10
101
38
0.102
No. of
patients
0.0097
0.0006
2327
No. of
patients
3-Year
survival
rate (%)
P
value
Increase
14
0.038
Decrease
95
38
\25%a
69
C25%
86
37
\50%a
23
43
C50%
72
38
\75%a
44
36
C75%
51
41
0.582
0.495
0.825
TABLE 6 Multivariate overall survival analysis of significant factors among previous univariate overall survival analysis
Factors
Hazard ratio
95% CI
P value
Head
1.34
0.912.09
0.138
Body or tail
1.0
1.162.17
0.004
0.891.69
0.198
0.912.68
0.115
0.851.72
0.305
1.182.28
0.004
0.912.68
0.561
Tumor location
Adjuvant chemotherapy
Yes
1.0
No
1.59
1.23
UICC pT factor
T1/2
1.0
T3
1.49
1.20
No
1.0
Postoperative CA19-9
\37
1.0
C37
1.64
1.13
Decrease
1.0
CI confidence interval
significant effect on overall survival of preoperative CA199 levels despite an adequate difference in 3-year survival
could be due to the smaller sample size. Moreover, the
preoperative CA19-9 cutoff values varied (from undetectable to 2,000 U/ml) among previous studies.15,17,18,24,30,33
Use of a single unique preoperative CA19-9 cutoff value
for postoperative prognosis seems to be difficult in patients
with resectable pancreatic cancer.
In contrast, postoperative CA19-9 levels were significantly higher in patients with R1, and with lymph node
metastasis. This result may reflect the microscopically
residual cancer cell after operation. In addition, differences
were significant for all four postoperative CA19-9 cutoff
values in this study. This result was similar to previously
reported findings, which demonstrated that lower postoperative CA19-9 values are associated with longer
survivals.16,17,19,32 In some previous reports, the most
significant prognostic cutoff value is defined as the cutoff
value with the lowest P value.17,18 In a similar method, 37
U/ml was the lowest P value among the four postoperative
CA19-9 cutoff points in our study, and it seemed to be most
useful for predicting long-term survival. Thus, for multivariate analysis of overall survival, we adopted 37 U/ml as
the postoperative CA19-9 cutoff point, which is based on
the standard deviation in a normal population.
The cutoff value for prediction of overall survival was
37 U/ml in our analysis, less than the 200 U/ml in the
analysis by Ferrone et al. and 180 U/ml in that by Montgomery et al.17,33 We suggested these differences might be
due to differences in the timing of postoperative measurement. Ferrone et al. considered that variability in the
postoperative CA19-9 cutoff point was a confounding
factor in their retrospective study.17 The choice of cutoff
point could lead to bias if the timing of the evaluation
preceded the decline in the patients health. Because the
same confounding factor certainly exists in our study, we
attempted to carry out CA19-9 measurement within a
narrow timeframe (approximately 1 month after surgery;
median, 32 days; range, 1457 days), and we consider that
the timing of measurement could minimize the influence of
adjuvant chemotherapy, which started between 2 and
6 weeks after surgery. Therefore, the decrease or normalization of postoperative CA19-9 seems to be due to the
efficacy of surgical resection alone. We considered that if
all carcinoma cells were eliminated by curative surgery,
postoperative CA19-9 level would normalize within a few
weeks, because the half-life serum CA19-9 level is
approximately 14 h.39 On the other hand, patients with
CA19-9 level outside normal limits at approximately
1 month after the operation might have had residual tumor
of adenocarcinoma, and they seemed to have a poor
prognosis. Actually, the 3-year survival rate for the [37 U/
ml postoperative CA19-9 group was only 9.5%. We concluded that postoperative CA19-9 is a better prognostic
marker than preoperative CA19-9 for patients with
resectable pancreatic cancer, because the difference in
overall survival was significant at all four cutoffs for postbut not preoperative CA19-9 level.
Moreover, patients with pre- to postoperative increase in
CA19-9 level had significantly poorer prognosis in this
study. The median number of days between pre- and
2328
N. Kondo et al.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
of recurrence after postoperative high-dose radiotherapy with 5FU-based concurrent chemotherapy for resectable pancreatic
cancer. Am J Clin Oncol. 2008;31:44653.
Murakami Y, Uemura K, Ohge H, Hayashidani Y, Sudo T, Sueda
T. Intraductal papillary-mucinous neoplasms and mucinous cystic
neoplasms of the pancreas differentiated by ovarian-type stroma.
Surgery. 2006;140:44853.
Steinberg W. The clinical utility of the CA 19-9 tumor-associated
antigen. Am J Gastroenterol. 1990;85:3505.
Tempero MA, Uchida E, Takasaki H, Burnett DA, Steplewski
Z, Pour PM. Relationship of carbohydrate antigen 19-9 and
Lewis antigens in pancreatic cancer. Cancer Res. 1987;47:
55013.
(UICC) International Union Against Cancer. TMN classification
of malignant tumors. 6th ed. New York: Wiley-Liss; 2002.
Waraya M, Yamashita K, Katagiri H, et al. Preoperative serum
CA19-9 and dissected peripancreatic tissue margin as determiners
of long-term survival in pancreatic cancer. Ann Surg Oncol.
2009;16:123140.
Fujioka S, Misawa T, Okamoto T, et al. Preoperative serum
carcinoembryonic antigen and carbohydrate antigen 19-9 levels
for the evaluation of curability and resectability in patients with
pancreatic adenocarcinoma. J Hepatobiliary Pancreat Surg.
2007;14:53944.
Ong SL, Garcea G, Thomasset SC, et al. Surrogate markers of
resectability in patients undergoing exploration of potentially
resectable pancreatic adenocarcinoma. J Gastrointest Surg.
2008;12:106873.
Smith RA, Bosonnet L, Ghaneh P, et al. The platelet-lymphocyte
ratio improves the predictive value of serum CA19-9 levels in
determining patient selection for staging laparoscopy in suspected
periampullary cancer. Surgery. 2008;143:65866.
Ko AH, Hwang J, Venook AP, Abbruzzese JL, Bergsland EK,
Tempero MA. Serum CA19-9 response as a surrogate for clinical
outcome in patients receiving fixed-dose rate gemcitabine for
advanced pancreatic cancer. Br J Cancer. 2005;93:1959.
2329
29. Wong D, Ko AH, Hwang J, Venook AP, Bergsland EK, Tempero
MA. Serum CA19-9 decline compared to radiographic response
as a surrogate for clinical outcomes in patients with metastatic
pancreatic cancer receiving chemotherapy. Pancreas. 2008;37:
26974.
30. Lundin J, Roberts PJ, Kuusela P, Haglund C. The prognostic
value of preoperative serum levels of CA 19-9 and CEA in
patients with pancreatic cancer. Br J Cancer. 1994;69:5159.
31. Nakao A, Oshima K, Nomoto S, et al. Clinical usefulness of CA19-9 in pancreatic carcinoma. Semin Surg Oncol. 1998;15:1522.
32. Beretta E, Malesci A, Zerbi A, et al. Serum CA 19-9 in the
postsurgical follow-up of patients with pancreatic cancer. Cancer.
1987;60:242831.
33. Montgomery RC, Hoffman JP, Riley LB, Rogatko A, Ridge JA,
Eisenberg BL. Prediction of recurrence and survival by postresection CA 19-9 values in patients with adenocarcinoma of the
pancreas. Ann Surg Oncol. 1997;4:5516.
34. Ni XG, Bai XF, Mao YL, et al. The clinical value of serum CEA,
CA19-9, and CA242 in the diagnosis and prognosis of pancreatic
cancer. Eur J Surg Oncol. 2005;31:1649.
35. Zhang S, Wang YM, Sun CD, Lu Y, Wu LQ. Clinical value of
serum CA19-9 levels in evaluating resectability of pancreatic
carcinoma. World J Gastroenterol. 2008;14:37503.
36. Maithel SK, Maloney S, Winston C, et al. Preoperative CA 19-9
and the yield of staging laparoscopy in patients with radiographically resectable pancreatic adenocarcinoma. Ann Surg
Oncol. 2008;15:351220.
37. Schlieman MG, Ho HS, Bold RJ (2003) Utility of tumor markers
in determining resectability of pancreatic cancer. Arch Surg
138:9515; discussion 56.
38. Safi F, Schlosser W, Falkenreck S, Beger HG. Prognostic value of
CA 19-9 serum course in pancreatic cancer. Hepatogastroenterology. 1998;45:2539.
39. Yoshimasu T, Maebeya S, Suzuma T, et al. Disappearance curves
for tumor markers after resection of intrathoracic malignancies.
Int J Biol Markers. 1999;14:99105.