Ann Surg Oncol (2010) 17:23212329

DOI 10.1245/s10434-010-1033-0

ORIGINAL ARTICLE PANCREATIC TUMORS

Prognostic Impact of Perioperative Serum CA 19-9 Levels

in Patients with Resectable Pancreatic Cancer
Naru Kondo, MD, Yoshiaki Murakami, MD, Kenichiro Uemura, MD, Yasuo Hayashidani, MD, Takeshi Sudo, MD,
Yasushi Hashimoto, MD, Akira Nakashima, MD, Ryutaro Sakabe, MD, Norifumi Shigemoto, MD,
Yasushi Kato, MD, Hiroki Ohge, MD, and Taijiro Sueda, MD
Department of Surgery, Division of Clinical Medical Science, Graduate School of Biomedical Sciences, Hiroshima
University, Hiroshima, Japan

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

Disclosure of commercial interest: None of the authors have any

commercial interest in the subject of this study and received any
financial or material support for this study.
Society of Surgical Oncology 2010
First Received: 17 December 2009;
Published Online: 25 March 2010
N. Kondo, MD
e-mail: k-naru@par.odn.ne.jp

analysis, postoperative adjuvant chemotherapy (hazard

ratio, 1.59; P = 0.004) and postoperative CA19-9 cutoff
value of 37 U/ml (HR, 1.64; P = 0.004) remained independent predictors of prognosis.
Conclusions. Postoperative CA19-9 level is a better
prognostic factor than preoperative CA19-9 level, and
curative surgery for resectable pancreatic cancer should be
tried regardless of the preoperative CA19-9 level.

Pancreatic cancer is one of the deadliest cancers, despite

recent improvements in surgical and postoperative management. Surgery remains the only curative treatment for
this devastating disease. However, the 5-year survival rate
of patients with potentially resectable pancreatic cancer
was reported not to exceed 20%.16 Recurrence and distant
metastasis occurs frequently in these patients even if
curative resection is performed. Therefore, surgery alone is
not sufficient for improving the prognosis of patients with
pancreatic cancer,7 and adjuvant chemotherapy or radiation
therapy or both is recommended.814 Discovering a surrogate marker of predicting prognostic significance would be
helpful for identifying patients with high risk of recurrence
and distant metastasis.
Recently, several studies revealed that perioperative
carbohydrate antigen 19-9 (CA19-9) level, which is a
tumor-associated antigen requiring expression of a sialylated Lewis blood group antigen to be expressed, is a useful
marker for predicting prognosis after resection.1519 However, preoperative CA19-9 level was useful according to
some reports, and postoperative CA19-9 level was useful
according to others. Moreover, the useful cutoff value of
CA19-9 varied among these reports. The purpose of this
study was to determine which was the more useful predictive factorpreoperative or postoperative CA19-9

2322

level. For that purpose, pre- and postoperative CA19-9

levels were measured in more than 100 patients with
pancreatic cancer at a single institution, and their relationship to clinicopathological factors and overall survival
were analyzed by univariate and multivariate methods.
PATIENTS AND METHODS

N. Kondo et al.

The regimen of adjuvant chemotherapy was reported

previously.13 Briefly, postsurgical adjuvant chemotherapy
(10 cycles, 1 cycle every 2 weeks, of intravenous gemcitabine [700 mg/m2 on day 1] plus orally administered
S-1 [50 mg/m2 for 7 consecutive days] followed by a
1-week pause of chemotherapy) was started between 2
and 6 weeks after surgery. No patient was exposed to
external beam radiation or intraoperative irradiation.

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

Serum CA19-9 Levels in Pancreatic Cancer

Against Cancer (UICC) tumor-node-metastasis (TMN)

classification.23

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.

Of the 109 patients, 56 (51%) were men and 53 (49%)

were women, with a median age of 70 (range, 4387)
years. The pancreatic tumor was confined to the head,
body, or tail of the pancreas in 78 (72%) and 31 patients
(28%), respectively. PD and PPPD were performed in 4
(3%) and 74 patients (68%), respectively, and DP and TP
in 29 (27%) and 2 patients (2%), respectively. Seventyseven (71%) and 32 (29%) patients had R0 and R1 resection, respectively. Tumors were identified as welldifferentiated adenocarcinoma in 68 patients (62%), moderately differentiated adenocarcinoma in 33 patients (31%),
and poorly differentiated adenocarcinoma in 8 patients
(7%). According to TMN classification, 9 (8%), 9 (8%),
and 91 patients (84%) had T1, T2, and T3 tumors,
respectively, and 69 patients (63%) had lymph nodes
metastasis. Finally, 5 (5%), 6 (5%), 26 (24%), 58 (53%),
and 14 (13%) patients were diagnosed with stage IA, IB,
IIA, IIB, and IV, respectively. All 14 patients with stage IV
disease had para-aortic lymph nodes metastasis detectable
only on postoperative histological examination, but not on
preoperative imaging examination (Table 1).
Relationship of Pre- and Postoperative CA19-9 Levels
to Clinicopathological Factors
The median time from preoperative CA19-9 measurement to surgery was 14 (range, 144) days. However, total
bilirubin level of seven patients was [4.0 mg/dl just before
surgery. The median time from surgery to postoperative
CA19-9 measurement was 33 (range, 1457) days. Pre- and
postoperative serum CA19-9 levels were compared for
each clinical and pathological factor (Table 2). Preoperative CA19-9 levels were significantly higher in patients
with UICC pT3 than in patients with UICC pT1/T2
(P = 0.022). There was no significant relationship to other
clinicopathological factors. In contrast, postoperative
CA19-9 levels were significantly higher in patients with R1
compared with those with R0 (P = 0.041), and in patients
with lymph node metastasis compared with those without
lymph node metastasis (P = 0.006).
Univariate Analysis of Overall Survival
The median follow-up time after surgery was 33 (range,
2150) months for 109 patients. Their overall survival rate
was 78% at 1 year, 55% at 2 years, and 28% at 5 years.
The log-rank test results of nine clinicopathological factors
are shown in Table 3. Three factors (gender, age, and
degree of pathological differentiation) did not influence

2324

N. Kondo et al.

TABLE 1 Characteristics and pathological assessments of 109

patients treated by resection of pancreatic cancer
No. of patients (%)
Age (yr)
Median

70

Range

4387

Gender
Male

56 (51)

Female

53 (49)

Tumor location
Head

78 (72)

Body or tail

31 (28)

among the four postoperative CA19-9 cutoff points, it

seemed to be most useful for predicting long-term survival
and was adopted as the postoperative CA19-9 cutoff value
for multivariate analysis of overall survival in this study.
As shown in Fig. 1, the Kaplan-Meier survival curves for
groups divided on the basis of the 37 U/ml cutoff value
revealed a significant difference, as previously described
(P = 0.0003).
As shown in Table 5, the pre- to postoperative increase
in CA19-9 level was significantly associated with poor
prognosis (P = 0.038); however, postoperative decrease
(whether 25%, 50%, or 75%) had no effect on overall
survival.

Type of resection
PD

4 (3)

PPPD

74 (68)

DP

29 (27)

TP

2 (2)

Adjuvant chemotherapy
Yes

81 (74)

No

28 (26)

Residual tumor (R factor)

R0

77 (71)

R1

32 (29)

Pathological differentiation
Well
Moderately

68 (62)
33 (31)

Poorly

8 (7)

Multivariate Analysis of Overall Survival

Finally, a multivariate proportional hazards regression
model was used to analyze seven factors identified by
previous univariate analysis (Table 6). The UICC final
stage was not included in the multivariate analysis, even
though it was significant on univariate analysis, because the
UICC stage itself depends upon pT stage and lymph node
metastasis, and seems to be confounded with them. This
analysis identified absence of adjuvant chemotherapy
(hazard ratio [HR], 1.59: 95% confidence interval [CI],
1.162.17; P = 0.004) and postoperative CA19-9 cutoff
value C 37 (HR, 1.64; 95% CI, 1.182.28; P = 0.004) as
independent predictors of poor overall survival.

UICC pT stage
T1

9 (8)

T2

9 (8)

T3

91 (84)

Lymph node metastasis

Yes

69 (63)

No

40 (37)

UICC final stage

IA

5 (5)

IB

6 (5)

IIA

26 (24)

IIB

58 (53)

IV

14 (13)

PD pancreatoduodenectomy, PPPD pylorus-preserving pancreatoduodenectomy, DP distal pancreatectomy, TP total pancreatectomy

overall survival. However, tumor location (P = 0.019),

postoperative adjuvant chemotherapy (P \ 0.001), R factor
(P \ 0.001), UICC pT stage (P = 0.004), lymph node
metastases (P = 0.015), and UICC final stage (P = 0.015)
were significantly associated with overall survival.
As shown in Table 4, postoperative CA19-9 level but
not preoperative CA19-9 level had a significant effect on
overall survival. Because 37 U/ml was the lowest P value

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

Serum CA19-9 Levels in Pancreatic Cancer

TABLE 2 Comparison of
serum perioperative CA19-9
levels in 109 pancreatic cancer
patients treated by resection and
divided into groups based on
clinicopathological factors

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

independent predictor for overall survival according to

Montgomery et al.33 Hence, to clarify whether preoperative
and/or postoperative CA19-9 is useful for prognosis of
patients with resectable pancreatic cancer, we determined
the serum preoperative CA19-9 level, serum postoperative
CA19-9 level, and pre- to postoperative change in levels
and assessed the relationship of these factors to overall
survival by univariate and multivariate analyses.
The Kaplan-Meier method and log-rank test for overall
survival was performed between groups divided on the
basis of four pre- and postoperative CA19-9 cutoff levels,
which had been previously reported.1518,24,26,27,3336
There was no significant difference in overall survival
between groups at all four preoperative CA19-9 cutoff
values, suggesting that preoperative CA19-9 level is not a

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.

TABLE 3 Univariate overall survival analysis of clinicopathological

factors for 109 patients treated by surgical resection for pancreatic
cancer
No. of
patients

3-Year survival rate

(%)

P
value

Overall
Survival
1.0

Postoperative CA19-9 < 37

Postoperative CA19 9 37

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

Lymph node metastasis

Yes
No

0.019

69
40

24
51

0.015

11

79

0.015

II/IV

98

27

UICC final stage

Number
at Risk
68
41

13
2

10
0

6
0

Years
44
22

30
4

FIG. 1 Comparison of postoperative overall survival between two

groups separated by the postoperative CA19-9 cutoff value of 37 U/
ml. A significant difference in survival was revealed by log-rank test
(P = 0.0003)

level is extremely high, as long as the tumor is judged to be

resectable by radiographic and other means.
Many previous studies have demonstrated that high
serum CA19-9 level is a significant factor for poor prognosis in pancreatic cancer.15,17,18,24,30,33 Although the
reasons for lack of a significant prognostic difference at all
four preoperative CA19-9 levels in our data were unclear,
one of them might be the difference in the distribution of
preoperative CA19-9 levels. The percentage of patients
with preoperative CA19-9 levels [1,000 U/ml in our data
(11.9%) seemed to be less than that reported by Ferrone
et al. (18.9%) or Waraya et al. (19.1%).17,24 Another possible reason might be sample size of our study. No

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

Serum CA19-9 Levels in Pancreatic Cancer

2327

TABLE 5 Univariate overall survival analysis at several levels of

pre- to postoperative CA19-9 change for 109 patients treated by
surgical resection for pancreatic cancer
Percentage change
from pre- to
postoperative CA19-9

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

14 patients with pre- to postoperative increase in CA19-9 level were

excluded

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

Residual tumor (R factor)

R0
1.0
R1

1.23

UICC pT factor
T1/2

1.0

T3

1.49

Lymph node metastasis

Yes

1.20

No

1.0

Postoperative CA19-9
\37

1.0

C37

1.64

Change of pre- to postoperative CA19-9

Increase

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

postoperative CA19-9 measurement was 48 (range, 2178).

We consider that the increase of CA19-9 reflects rapid
regrowth of the cancer after surgical resection, and patients
with such an increase frequently had very poor prognosis.
However, postoperative CA19-9 level more than 37 U/ml
was the only independent perioperative CA19-9-related
factor for poor prognosis identified in multivariate overall
survival analysis. Therefore, normalization of postoperative CA19-9 level was more important for longer survival
than was pre- to postoperative decrease in CA19-9.
The effectiveness of postoperative adjuvant chemotherapy in patients with resectable pancreatic cancer has
been demonstrated in several randomized, controlled trials.810 In our study, use of postoperative adjuvant
chemotherapy with gemcitabine and S-1 also was an
independent factor for longer postoperative survival in
multivariate overall survival analysis. Previous studies also
have reported that the combination of gemcitabine and S-1
improves the prognosis of patients with pancreatic cancer.1214 Consequently, we emphasize the indispensability
of using postoperative adjuvant chemotherapy to increase
survival after surgical resection of pancreatic cancer, and
the importance of a surrogate marker of prognostic significance to identify patients at high risk for developing
recurrence and distant metastasis. Postoperative CA19-9
could be the candidate marker. We suggest that adjuvant
chemotherapy should be undertaken regardless the postoperative CA 19-9 levels, nevertheless, patients with
postoperative CA 19-9 levels [37 would have poor
prognosis.
Thus, we believe that postoperative CA19-9 level may
be clinically useful for postoperative prognostic evaluation
in cases of resectable pancreatic cancer, because postoperative treatment can be individualized based on its value.
However, this is a retrospective study over 10 years with
109 patients. Prospective validation, including an adequate
number of patients, is needed to clarify the relationship
between postoperative CA19-9 value and prognosis.
CONCLUSIONS
The present study demonstrated that postoperative
CA19-9 is a better prognostic marker than preoperative
CA19-9, and curative resection should be performed
regardless of the preoperative CA19-9 level. Moreover,
postoperative CA19-9 value [37 U/ml was shown to be an
independent significant prognostic factor for identifying
patients with poor prognosis.
REFERENCES
1. Winter JM, Cameron JL, Campbell KA, et al. 1423 pancreaticoduodenectomies for pancreatic cancer: a single-institution

N. Kondo et al.

2.

3.

4.

5.

6.

7.
8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

experience. J Gastrointest Surg. 2006;10:1199210; discussion

2101.
Shimada K, Sakamoto Y, Sano T, Kosuge T. Prognostic factors
after distal pancreatectomy with extended lymphadenectomy for
invasive pancreatic adenocarcinoma of the body and tail. Surgery. 2006;139:28895.
Moon HJ, An JY, Heo JS, Choi SH, Joh JW, Kim YI. Predicting
survival after surgical resection for pancreatic ductal adenocarcinoma. Pancreas. 2006;32:3743.
Brown KM, Domin C, Aranha GV, Yong S, Shoup M. Increased
preoperative platelet count is associated with decreased survival
after resection for adenocarcinoma of the pancreas. Am J Surg.
2005;189:27882.
Wagner M, Redaelli C, Lietz M, Seiler CA, Friess H, Buchler
MW. Curative resection is the single most important factor
determining outcome in patients with pancreatic adenocarcinoma.
Br J Surg. 2004;91:58694.
Kuhlmann KF, de Castro SM, Wesseling JG, et al. Surgical
treatment of pancreatic adenocarcinoma; actual survival and
prognostic factors in 343 patients. Eur J Cancer. 2004;40:549
58.
Traverso LW. Pancreatic cancer: surgery alone is not sufficient.
Surg Endosc. 2006;20(Suppl 2):S4469.
Oettle H, Post S, Neuhaus P, et al. Adjuvant chemotherapy with
gemcitabine vs observation in patients undergoing curative-intent
resection of pancreatic cancer: a randomized controlled trial.
JAMA. 2007;297:26777.
Neoptolemos JP, Stocken DD, Friess H, et al. A randomized trial
of chemoradiotherapy and chemotherapy after resection of pancreatic cancer. N Engl J Med. 2004;350:120010.
Kosuge T, Kiuchi T, Mukai K, Kakizoe T. A multicenter randomized controlled trial to evaluate the effect of adjuvant
cisplatin and 5-fluorouracil therapy after curative resection in
cases of pancreatic cancer. Jpn J Clin Oncol. 2006;36:15965.
Takada T, Amano H, Yasuda H, et al. Is postoperative adjuvant
chemotherapy useful for gallbladder carcinoma? A phase III
multicenter prospective randomized controlled trial in patients
with resected pancreaticobiliary carcinoma. Cancer. 2002;95:
168595.
Murakami Y, Uemura K, Sudo T, et al. Impact of adjuvant
gemcitabine plus S-1 chemotherapy after surgical resection for
adenocarcinoma of the body or tail of the pancreas. J Gastrointest
Surg. 2009;13:8592.
Murakami Y, Uemura K, Sudo T, et al. Adjuvant gemcitabine
plus S-1 chemotherapy after surgical resection for pancreatic
adenocarcinoma. Am J Surg. 2008;195:75762.
Murakami Y, Uemura K, Sudo T, et al. Postoperative adjuvant
chemotherapy improves survival after surgical resection for
pancreatic carcinoma. J Gastrointest Surg. 2008;12:53441.
Berger AC, Meszoely IM, Ross EA, Watson JC, Hoffman JP.
Undetectable preoperative levels of serum CA 19-9 correlate with
improved survival for patients with resectable pancreatic adenocarcinoma. Ann Surg Oncol. 2004;11:6449.
Berger AC, Garcia M Jr, Hoffman JP, et al. Postresection CA 199 predicts overall survival in patients with pancreatic cancer
treated with adjuvant chemoradiation: a prospective validation by
RTOG 9704. J Clin Oncol. 2008;26:591822.
Ferrone CR, Finkelstein DM, Thayer SP, Muzikansky A, Fernandez-del Castillo C, Warshaw AL. Perioperative CA19-9 levels
can predict stage and survival in patients with resectable pancreatic adenocarcinoma. J Clin Oncol. 2006;24:2897902.
Kang CM, Kim JY, Choi GH, et al. The use of adjusted preoperative CA 19-9 to predict the recurrence of resectable pancreatic
cancer. J Surg Res. 2007;140:315.
Kinsella TJ, Seo Y, Willis J, et al. The impact of resection margin
status and postoperative CA19-9 levels on survival and patterns

Serum CA19-9 Levels in Pancreatic Cancer

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.