Journal of Surgical Oncology 2014;109:198201

Appropriate Gastrectomy Resection Margins for Early Gastric Carcinoma

BEOM SU KIM, MD, SEONG TAE OH, MD, JEONG HWAN YOOK, MD, HEE SUNG KIM, MD,
IN SEOB LEE, MD, AND BYUNG SIK KIM, MD, PhD*
Department of Surgery, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea

Objective: In Korea and Japan, early gastric cancer (EGC) accounts for >50% of all gastric cancers. Here, we propose recommendations for the
optimal distance from the tumor to the resection margins when evaluating EGC.
Summary of Background Data: There are very few guidelines regarding the distance from the EGC tumor to the resection margins.
Methods: We evaluated 2,081 patients who underwent gastrectomy for EGC between January 1989 and May 2000. We subdivided tumors according
to the distance from the proximal margin:
1, >1,
10, >10,
30, or >30 mm.
Results: Three of ve patients demonstrating distances
1 mm between the tumor and gross proximal margin were microscopically positive. No
patients with gross proximal margins >1,
10, >10, or
30 mm were microscopically positive. There were no statistical differences in rates of
microscopically positive margin, reresection, or reoperation between groups (P > 0.05). In addition, there were statistical differences in terms of
tumor recurrence and diseaserelated death between groups (P > 0.05).
Conclusions: When the resection margins are clear, we propose that margins >1 mm are adequate for EGC gastrectomy.

J. Surg. Oncol. 2014;109:198201. 2013 Wiley Periodicals, Inc.

KEY WORDS: early gastric cancer; resection margin; reresection

INTRODUCTION
Early gastric cancer (EGC) is dened by the Japanese Society of
Gastroenterological Endoscopy as conned to either the mucosa or
submucosa, irrespective of lymph node metastasis [1]. In Korea and
Japan, >50% of all gastric cancers are EGC [2,3]. However, there
are very few published guidelines on the resection margins for EGC.
Based on a Medline search, only two articles recommend resection
margins [4,5]. Bozzetti et al. evaluated 343 patients with gastric cancer
(including advanced cancer), specically referencing the distance from
the outline of the grossly assessed tumor and other macroscopic and
microscopic features. They reported that proximal and distal inltration
distances >3 cm did not occur in lesions conned to the mucosa,
submucosa, or muscularis [5]. However, that study was reported
30 years ago and may not accurately reect gastric cancer today.
According to the 2010 Japanese gastric cancer treatment guidelines
(version 3), gross resection margins 2 cm should be obtained for T1
tumors [4]. However, these guidelines do not cite any clinical studies,
making it difcult to assess reliability.
In our current study, we evaluate the incidence of positive resection
margins on microscopy and tumor recurrence, taking into account the
gross distance from the tumor to the resection margins. The purpose of
this study was to recommend the appropriate gross distances from the
tumor to the resection margins in EGC patients.

classied as differentiated or undifferentiated adenocarcinomas. We

evaluated the distance from the gross resection margin and
microscopically positive margins. Tumors that were grossly abutting
or a very short distance from the resection margin were dened as

1 mm. We evaluated the longterm followup results of each patient to
determine any associations between tumor recurrence, diseaserelated
death, diseaseunrelated death, and gross distance from the tumor to the
resection margin.
We included patients who underwent curative EGC resection and
excluded patients with Siewert type I and II adenocarcinoma of the
esophagogastric junction [7]. Siewert type I adenocarcinoma aficts
the distal esophagus, and type II is the true carcinoma of the cardia.
Patients with tumor centers located <1cm oral or <2cm aboral to the
esophagogastric junction were also excluded. In type I and II cancers,
the real proximal margin can be quite different from the segment of the
anvil, following total gastrectomy.
All statistical analysis was performed using the Statistical Package
for Social Science (SPSS version 12.0 for Windows; SPSS, Inc.,
Chicago, IL). Oneway analysis of variance (oneway ANOVA) and
the Tukeys method were used for multiple comparisons of the
mean data. The chisquared test was used to identify the risk factors
associated with tumor recurrence and lymph node metastasis. The
KaplanMeier method and logrank test was used to analyze the survival
rate and tumor recurrence. A Pvalue < 0.05 indicates statistical
signicance.

PATIENTS AND METHODS

We retrospectively reviewed 2,081 EGC patients who underwent
gastrectomy at the Asan Medical Center, University of Ulsan College of
Medicine, Seoul, Korea between January 1989 and May 2000. We also
wanted to determine the longterm owup outcomes over this 10year
period. We analyzed the clinicopathologic characteristics according to
the Japanese classications for gastric cancer [6]. Tumors were classied
as supercial, protruded, or excavated according to the macroscopic
classications for gastric cancer by the Japanese Gastric Cancer
Association [6]. Tumor sites were equally divided into the upper,
middle, and lowerthird sections of the stomach, and tumors were

2013 Wiley Periodicals, Inc.

*Correspondence to: Correspondence to: Byung Sik Kim, MD, PhD, 3881,
Pungnap 2 dong, SongpaGu, Seoul 138736, Korea. Fax: 8224749027.
Email: bskim@amc.seoul.kr
Received 10 May 2013; Accepted 10 October 2013
DOI 10.1002/jso.23483
Published online 19 November 2013 in Wiley Online Library
(wileyonlinelibrary.com).

Resection Margins in Early Gastric Cancer

199

RESULTS

or reoperation (Table II; P > 0.05). In addition, there were no statistical

differences in tumor recurrence or diseaserelated death (Fig. 1).

The general clinicopathological characteristics of the study cohort are

summarized in Table I. The median followup period was 128.2 months
(range: 0.2265.8). The median number of retrieved lymph nodes was
25 (range: 586).

Surgical Outcomes and Tumor Recurrence in Patients

With Distances
1 mm

Clinicopathologic Characteristics and Outcomes According to

the Gross Distance From the Tumor to the Resection Margin

Surgical outcomes and tumor recurrence are summarized in Table III.

In total, ve patients demonstrated abutting tumors that were very close
to the proximal resection margin. Three of these ve patients (75%)
demonstrated microscopically positive margins and underwent
reresection after primary surgery.

We subdivided tumors according to the distance from the tumor to the

proximal margin:
1 mm (very close or abutting), >1,
10, <10,
30,
or >30 mm. The clinicopathologic characteristics according to gross
distance from the tumor to the resection margin are summarized in
Table II. Five patients with distances
1 mm are separately summarized
in Table III because we could not statistically analyze this small number
of patients. Regarding the proximal margin, there were statistically
signicant differences between groups in terms of tumor location,
operation type, and tumor size, but no signicant differences in sex,
depth of invasion, macroscopic ndings, or lymph node metastasis.
There were differences in operation type and tumor location between
groups because total gastrectomy was primarily performed on the
patients with tumors located in the upperthird section and subtotal
gastrectomy was primarily performed on patients with tumors located in
the mid or lowerthird section of the stomach. There were no statistical
differences in the rates of microscopically positive margins, reresection,

TABLE I. Clinicopathologic Characteristics

Characteristics
Age (years)
Sex
Male
Female
Operation type
Subtotal gastrectomy
Total gastrectomy
Tumor location
Lower third
Middle third
Upper third
Tumor size (mm)
Distant from proximal resection margin (mm)
Depth of invasion
Mucosa
Submucosa
Macroscopic finding
Superficial
Protruded
Excavated
Histological type
Differentiated
Undifferentiated
Lymph node metastasis
No
Yes
Lymphatic/venous invasion
No
Yes
Tumor recurrence
No
Yes
Multiple cancers
No
Yes
No., number; SD, standard deviation.

Journal of Surgical Oncology

No.
2,081

Percent
100

1,374
707

66.0
34.0

1,893
188

91.0
9.0

1,306
634
141
2,081
2,081

62.8
30.5
6.8
100
100

1,039
1,042

49.9
50.1

1,744
120
217

83.8
5.8
10.4

1,077
1,024

51.8
48.2

1,827
254

87.8
12.2

1,884
197

90.5
9.5

1,983
98

95.3
4.7

2,010
71

96.6
3.4

Mean  SD
54.9  11.5

30.3  18.8
52.3  27.5

Outcomes in Patients With Microscopically

Positive Resection Margins
Table IV shows the outcomes of ve patients with microscopically
visible tumor invasion into the proximal resection margin. Two of the
patients with a positive proximal margin had tumor 3040 mm from the
proximal resection margin; however, these two patients demonstrated
unclear tumor borders and tumor margins that could not be clearly
identied at the time of surgery. Therefore, a signicant amount of tumor
remained in the remnant stomach after primary surgery, and these
patients ultimately underwent total gastrectomy. Three patients with
microscopically visible tumor invasion demonstrated proximal resection
margins <1 mm (i.e., very close to the resection margin).

DISCUSSION
The negative effect of positive resection margin on patient survival
has been the subject of many studies [814]. Most such reports describe
a positive resection margin as a negative prognostic factor in patients
with EGC with or without lymph node metastasis [810,13,14].
However, a positive resection margin is not a prognostic factor in
patients with aggressivestage gastric cancer or lymph node
metastasis [8,11,14,15]. Therefore, multiple operations are required
by patients with nonaggressive gastric cancer in order to obtain
microscopically negative resection margins [811,13]. Sun et al. [15]
reported that patients with advanced cancer are more likely to die of
carcinomatosis or disseminated disease than anastomotic recurrence.
Sano and Mudan [11] recommends patients with >N3 nodal
involvement should not undergo multiple operations in order to
obtain microscopically negative margins.
Cho et al. [8] attributed their low rate of positive margins (98% R0
rate; determined using microscopy) to surgical expertise, preoperative
endoscopic assessment, and the routine examination of frozen sections.
In our present study, ve patients (0.23%) demonstrated microscopically
positive resection margins. However, frozen sections, which
demonstrate higher accuracy, were not used to assess two patients.
Cho et al. and the Japanese gastric cancer treatment guidelines
recommend preoperatively marking the tumor border using clips
when the tumor border is unclear; according to their biopsy results, this
helps determine the best resection line [4,15]. In fact, it is difcult to
identify tumor lesions through a small opening in the stomach during
operations when the tumor is small with an unclear border. In two of ve
patients with microscopic residual tumors, the operator did not identify a
denite tumor lesion during surgery. These two tumor lesions were
located in the remnant stomach, and total gastrectomies were
subsequently performed on these two patients. If these two patients
had received preoperative endoscopic evaluation with clipping at the
proximal border, a negative resection margin would have been obtained
and additional gastrectomies would have been avoided.
There are very few denitive guidelines in the clinical literature that
recommend distances from the tumor to the resection margin. Of 48
patients who had gross distances
10 mm to the proximal resection
margin, we identied three patients with microscopically positive
resection margins. These three patients underwent multiple resections

TABLE II. Clinicopathologic Characteristics and Outcomes According to Gross Distance From the Proximal Resection Margin
Distance from proximal resection margin (mm)
Characteristics
Age (years)
Sex
Male
Female
Depth of invasion
Mucosa
Submucosa
Macroscopic finding
Superficial
Protruded
Excavated
Retrieved lymph nodes (no.)
Lymph node metastasis
No
Yes
Microscopically positive margin
No
Yes
Reresection or reoperation
No
Yes
Tumor recurrence, n (%)
No
Yes
Diseaserelated death, n (%)

>1,
10 (n 43)

>10,
30 (n 500)

>30 (n 1,533)

Pvalue

53.7  14.3

53.5  12.1

55.4  11.2

NS
NS

28
15

324
176

1018
515

18
25

253
247

764
769

39
1
3
24.3  12.1

409
33
58
25.3  13.2

1291
86
156
24.4  12.5

36
7

45
56

1343
190

43
0

499
1

1632
1

43
0

499
1

1632
1

42
1 (2.3%)
2 (4.6%)

469
30 (6%)
26 (5.2%)

1467
67 (4.4%)
57 (3.7%)

NS

NS

NS
NS

NS

NS
NSa
NSa

no., number; NS, nonspecic.

a
Determined using KaplanMeier analysis.
TABLE III. Outcomes in Patients With Short Gross Distances (
1 mm) From the Proximal Resection Margin
Patient
1
2
3
4
5

Gross findings

Microscopically positive margin

Reoperation after surgery

Superficial flat
Superficial Elevated
Superficial flat
Superficial flat
Superficial flat

Yes
No
Yes
No
Yes

No
No
Yes
No
Yes

Tumor recurrence (f/u months)

No
No
No
No
No

(180)
(185)
(180)
(157)
(157)

f/u, followup.

Fig. 1. Tumor recurrence and diseaserelated survival curves determined using the KaplanMeier method. There are no signicant differences
between the resection margin subgroups (P > 0.05). A: >1 and
10 mm. B: >10 and
30 mm. C: >30 mm.
Journal of Surgical Oncology

Resection Margins in Early Gastric Cancer

201

TABLE IV. Outcomes in Patients With Microscopically Positive Margins

Characteristics

Gross distance from the resection margin (mm)

Positive proximal margin (n 5)

1
2
3
4
5

30
40

1

1

1

Gross findings

Reoperation after surgery

Superficial depressed
Superficial flat
Superficial depressed
Superficial flat
Superficial flat

Yes
Yes
Yes
Yes
No

Tumor recurrence (f/u months)

No
No
No
No
No

(187)
(186)
(180)
(180)
(157)

f/u, followup.

after the rst operation. In addition, all of these patients demonstrated

gross distances
1 mm to the resection margin. Patients with distances
>1 mm from proximal resection to the clear tumor margin demonstrated
neither microscopically visible tumor invasion into the resection margin
nor reresection.
There were no statistical differences in tumor recurrence or disease
related death between the three groups of grossly distance from resection
margin. All patients with microscopically visible tumor invasion into the
resection margin underwent reresection and no microscopically visible
tumors was subsequently noted the resection margins. This means that
the stomach should be grossly resected 1 mm proximal to the upper
border of the tumor in order to obtain microscopically negative EGC
margins.
We were only able to perform a limited evaluation of the gross
distance to the resection margins due to the segment of anvil or staple
line. In this study, a 25 or 28mm circular anvil (we preferred 25 mm)
was inserted into the esophagus and segment of anvil, which contains a
small portion of the esophagus and jejunum that are removed after
anastomosis. In fact, the real gross distance from the resection margins
could have been >2 mm following subtotal gastrectomy due to the staple
line (we usually resected the stomach using a linear stapler) and >5
6 mm following total gastrectomy due to the segment of anvil (we
usually resected the esophagus and anvil using a circular stapler).
However, differences between the grossly measured distance and the
real distance from the resection margin following total gastrectomy
might not be that great because we excluded patients with Siewert type I
and II adenocarcinoma of the esophagogastric junction.
In conclusion, when the resection margins are clear, a gross distance
>1 mm from the proximal resection is adequate for assessing EGC
following gastrectomy.

REFERENCES
1. Murakami T: Pathomorphological diagnosis: Denition and gross
classication of early gastric cancer. Gann Monogr Cancer Res
1971;11:5355.
2. Fujii M, Sasaki J, Nakajima T: State of the art in the treatment of
gastric cancer: From the 71st Japanese Gastric Cancer Congress.
Gastric Cancer 1999;2:151157.

Journal of Surgical Oncology

3. Kim C, Lee S, Yang D: What is the prognosis for early gastric cancer
with pN stage 2 or 3 at the time of preoperation and operation. J
Korean Gastric Cancer Assoc 2006;6:114119.
4. Japanese gastric cancer treatment guidelines 2010 (ver. 3). Gastric
Cancer 2011;14:113123.
5. Bozzetti F, Bonfanti G, Bufalino R, et al.: Adequacy of margins of
resection in gastrectomy for cancer. Ann Surg 1982;196:685690.
6. Japanese Gastric Cancer Association: Japanese Classication of
Gastric Carcinoma2nd English Edition. Gastric Cancer 1998;1:
1024.
7. Siewert JR, Stein HR: Carcinoma of the cardia: Carcinoma of the
gastroesophageal junctionClassication, pathology and extent of
resection. Dis Esophagus 1996;9:173182.
8. Cho BC, Jeung HC, Choi HJ, et al.: Prognostic impact of resection
margin involvement after extended (D2/D3) gastrectomy for
advanced gastric cancer: A 15year experience at a single institute.
J Surg Oncol 2007;95:461468.
9. Nagata T, Ichikawa D, Komatsu S, et al.: Prognostic impact of
microscopic positive margin in gastric cancer patients. J Surg Oncol
2011;104:592597.
10. Raziee HR, Cardoso R, Seevaratnam R, et al.: Systematic
review of the predictors of positive margins in gastric cancer
surgery and the effect on survival. Gastric Cancer 2012;15:S116
S124.
11. Sano T, Mudan SS: No advantage of reoperation for positive
resection margins in node positive gastric cancer patients? Jpn J Clin
Oncol 1999;29:283284.
12. Shiu MH, Moore E, Sanders M, et al.: Inuence of the extent of
resection on survival after curative treatment of gastric carcinoma. A
retrospective multivariate analysis. Arch Surg 1987;122:1347
1351.
13. Songun I, Bonenkamp JJ, Hermans J, et al.: Prognostic value
of resectionline involvement in patients undergoing curative
resections for gastric cancer. Eur J Cancer 1996;32A:433
437.
14. Wang SY, Yeh CN, Lee HL, et al.: Clinical impact of positive
surgical margin status on gastric cancer patients undergoing
gastrectomy. Ann Surg Oncol 2009;16:27382743.
15. Sun Z, Li DM, Wang ZN, et al.: Prognostic signicance of
microscopic positive margins for gastric cancer patients with
potentially curative resection. Ann Surg Oncol 2009;16:3028
3037.