PDF icon Download PDF
Endoscopy 2012; 44(06): 577-583
DOI: 10.1055/s-0031-1291712
Original article
© Georg Thieme Verlag KG Stuttgart · New York

Clinical outcomes of endoscopic submucosal dissection for early gastric cancer in remnant stomach or gastric tube

N. Nishide
Division of Endoscopy, Shizuoka Cancer Center, Shizuoka, Japan
,
H. Ono
Division of Endoscopy, Shizuoka Cancer Center, Shizuoka, Japan
,
N. Kakushima
Division of Endoscopy, Shizuoka Cancer Center, Shizuoka, Japan
,
K. Takizawa
Division of Endoscopy, Shizuoka Cancer Center, Shizuoka, Japan
,
M. Tanaka
Division of Endoscopy, Shizuoka Cancer Center, Shizuoka, Japan
,
H. Matsubayashi
Division of Endoscopy, Shizuoka Cancer Center, Shizuoka, Japan
,
Y. Yamaguchi
Division of Endoscopy, Shizuoka Cancer Center, Shizuoka, Japan
› Author Affiliations
Further Information

Corresponding author

N. Kakushima, MD, PhD
Division of Endoscopy
Shizuoka Cancer Center
1007 Shimonagakubo
Nagaizumi
Suntogun
Shizuoka, 411-8777
Japan   
Phone: +81-55-9895222   
Fax: +81-55-9895692   

Publication History

submitted 01 May 2011

accepted after revision 15 January 2012

Publication Date:
08 March 2012 (online)

Also available ateRef
 
 PDF Download Permissions and Reprints

Background and study aims: Little information exists regarding the optimal treatment of early gastric cancer (EGC) in a remnant stomach or gastric tube. The aim of this study was to assess the feasibility and clinical outcomes of endoscopic submucosal dissection (ESD) for EGC in a remnant stomach and gastric tube.

Patients and methods: Between September 2002 and December 2009, ESD was performed in 62 lesions in 59 patients with EGC in a remnant stomach (48 lesions) or gastric tube (14 lesions). Clinicopathological data were retrieved retrospectively to assess the en bloc resection rate, complications, and outcomes. Treatment results were assessed according to the indications for endoscopic resection, and were compared with those of ESD performed in a whole stomach during the same study period.

Results: The en bloc resection rates for lesions within the standard and expanded indication were 100 % and 93 %, respectively. Postoperative bleeding occurred in five patients (8 %). The perforation rate was significantly higher (18 %, 11 /62) than that of ESD in a whole stomach (5 %, 69 /1479). Among the perforation cases, eight lesions involved the anastomotic site or stump line, and ulcerative changes were observed in five lesions. The 3-year overall survival rate was 85 %, with eight deaths due to other causes and no deaths from gastric cancer.

Conclusion: A high en bloc resection rate was achieved by ESD for EGC in a remnant stomach or gastric tube; however, this procedure is still technically demanding due to the high complication rate of perforation.


 

Introduction

Little information exists regarding the optimal treatment of early gastric cancer (EGC) in a remnant stomach or a gastric tube. To obtain a radical cure, surgical resection has been considered the standard treatment; however, surgery is not usually preferred because of its high morbidity and mortality [1] [2]. Endoscopic mucosal resection (EMR) including endoscopic submucosal dissection (ESD) has been widely accepted as a standard treatment for EGC in Japan [3−5]. Recently, the benefit of endoscopic treatment for metachronous EGC in a reconstructed stomach has been reported in several case series of small numbers of patients [6−13]. Although ESD is a minor invasive treatment compared with surgery, performing ESD is considered technically more difficult in a remnant stomach or a gastric tube than in a whole stomach because of the deformity and narrow space due to previous surgery. Therefore, ESD for EGC in a remnant stomach or gastric tube is not yet widespread. In this study, we characterize the feasibility and clinical outcome of ESD for EGC in a remnant stomach or a gastric tube.


 

Patients and methods

From a prospectively entered database, which contained findings of 1904 EGC lesions that were treated by ESD at Shizuoka Cancer Center, between September 2002 and December 2009, 62 lesions in 59 patients with a remnant stomach after distal gastrectomy (48 lesions) or a gastric tube (14 lesions) were selected as the main study group for analysis. Among 1824 lesions treated by ESD for EGC in a whole stomach, 775 lesions within the standard criteria and 704 lesions within the expanded criteria of indications for ESD, as described below, were selected as a control group. Patients who underwent ESD for EGC in a stomach after partial gastrectomy (11 lesions), for EGC in a stomach after proximal gastrectomy (4 lesions), and patients who underwent combined therapy such as argon plasma coagulation and photodynamic therapy with ESD (3 lesions) were excluded ( [Fig.1]).

Zoom
Fig. 1 Flow chart showing the patient inclusion in the study.

The general indications for ESD at the Shizuoka Cancer Center are lesions with the finding of EGC that meet the criteria for EMR proposed by the Japanese Gastric Cancer Society [14]. The criteria include: those with a preoperative diagnosis of differentiated-type intramucosal cancer without ulcer findings; differentiated-type intramucosal cancer that is no larger than 3 cm in diameter with ulcer findings; differentiated-type minute invasive submucosal cancer (invasion less than 500 μm below the muscularis mucosa) that is no larger than 3 cm in diameter; and undifferentiated-type intramucosal cancer that is no larger than 2 cm in diameter without ulcer findings.

Method of ESD

As described previously by Ono et al. [5] [15], an insulated-tip diathermic knife (IT Knife2; KD-611 L, Olympus, Tokyo, Japan) was used as the main electrosurgical endoscopic knife. An endoscope with a water-jet function (GIF-Q260 J or GIF-2T260 M, Olympus, Tokyo, Japan) and a high frequency generator (VIO300 D, ERBE, Elektromedizin, Tübingen, Germany) was used. After recognizing the lesion by observation during chromoendoscopy with indigo carmine ([Fig. 2a, b]), dots marking the area around the lesion were placed using an argon plasma coagulation (APC) probe ([Fig. 2 c]). A commercially available solution of 0.4 % sodium hyaluronate (MucoUp; Johnson and Johnson ASP Japan, Tokyo, Japan) was injected through a 25-gauge needle into the submucosal layer under the lesion. A 1 – 2-mm precut was made with a needle knife followed by a circumferential mucosal incision around the lesion with the IT Knife2 ([Fig. 2 d, e]). The submucosal layer was then dissected by the IT Knife2 with an additional submucosal injection ([Fig. 2 f]). When the tissue was too hard to cut with the IT Knife2 because of the existence of fibrosis or staples from the previous surgery, a needle knife was used as appropriate. After the resected specimen had been retrieved, the ulcer bed was carefully examined for residual blood vessels and coagulated with hot biopsy forceps (Boston Scientific, Tokyo, Japan), IT Knife2, or APC ([Fig. 2 g, h]). Proton pump inhibitors were administered initially by injection and then orally after the procedure. A soft meal was started 2 days after ESD if there were no complications of postoperative bleeding or perforation.

Zoom
Fig. 2 The endoscopic submucosal dissection procedure. a A reddish, slightly depressed lesion with a small nodule is observed in a narrow area of a gastric tube. b Chromoendoscopy with indigo carmine is performed to determine the border of the lesion. c Marking dots were placed around the lesion using an argon plasma coagulation probe. d, e Circumferential cutting was performed with an insulated-tip diathermic knife (IT Knife2). f The presence of staples was revealed during submucosal dissection. g The artificial ulcer after removing the lesion is carefully observed for residual vessels. h The lesion proved to be a mucosal cancer (blue lines, number 3 – 8), and was removed en bloc with negative margins.
Zoom
Zoom
Zoom
Zoom
Zoom
Zoom
Zoom

 

Curability criteria from the histological assessment

The tumor size, depth of invasion, presence of ulcerative changes, existence of lymphovascular infiltration, and whether the specimen margins were free of cancer (R0) were assessed histopathologically. According to the criteria of the Japanese guidelines, the lesions were divided into one of three groups: standard indication, expanded criteria, and outside the criteria [14].

Lesions in the standard indication group consisted of differentiated-type mucosal cancer that was 2 cm or smaller in diameter, with no ulceration. The expanded criteria group consisted of differentiated-type mucosal cancer without ulceration; differentiated-type cancer that was 3 cm or smaller in diameter, with a depth of submucosal invasion less than 500 μm, with or without ulceration; or undifferentiated-type mucosal cancer that was 2 cm or smaller in diameter with no ulceration. Lesions that did not meet either the standard or the expanded criteria were defined as outside the criteria.

Curative resection was considered to be lesions that fitted the standard or expanded criteria with no lymphovascular infiltration, and that were resected en bloc with R0 resection.

For patients who were considered to have undergone curative resection for a lesion within the standard indication, follow-up by annual endoscopy with chromoendoscopy was performed. For patients who were considered to have undergone curative resection for a lesion within the expanded criteria, follow-up was performed by annual endoscopy with chromoendoscopy, and by abdominal computed tomography or abdominal ultrasonography. For patients who had undergone a non-curative resection, additional surgery was recommended.


 

Measurements

Operative time (from marking to complete lesion removal), en bloc resection rate, en bloc with R0 resection rate, complications (perforation and postoperative bleeding), and recurrence during the follow-up were retrieved for data analysis. Treatment results of ESD in a remnant stomach or gastric tube according to the indications for endoscopic resection were assessed, and compared with the results of ESD performed in a whole stomach. In this study, postoperative bleeding was defined as cases with hematemesis or melena requiring endoscopic intervention. Overall survival was measured from the date of ESD to death or to the confirmed date of the last follow-up visit. Survival time was calculated by the Kaplan – Meier method.

Statistical analysis was done using the Mann – Whitney’s U test and chi-squared test, to describe certain differences in an exploratory way, and a P value of < 0.05 was considered significant. For quantitative variables with apparently skewed distributions, the summary data were expressed as median (interquartile range [IQR] between the 25th and 75th percentiles). All analyses were done by StatView version 5.0 (SAS Institute Inc., Cary, North Carolina, USA). Written informed consent was obtained from all patients undergoing ESD. This retrospective study was endorsed by the institutional review board of our hospital (No. 23-J2 – 23 – 1-3).


 

 

Results

Overall clinicopathological characteristics of the 62 lesions in the 59 patients are summarized in  [Table 1]. Among 46 patients with a residual stomach, 29 patients (63 %) had undergone previous surgery for gastric cancer. All 13 patients with a gastric tube had undergone previous surgery for esophageal cancer. The median period between the first surgery and ESD was 30 years for patients who received distal gastrectomy for a benign disease, 6 years for patients who received distal gastrectomy for a gastric cancer, and 4.5 years for patients who received esophagectomy for an esophageal cancer. Most of the patients were male (56 /59, 95 %) and the lesions were mostly located in the upper (39 %) or middle (55 %) portion of the stomach. The median diameter of the tumors was 22 mm. Ulceration was present in 19 lesions (31 %). In 29 lesions (47 %) the anastomotic site or the stump line from the previous surgery were involved.

Table 1

Clinicopathological characteristics of the 62 lesions in 59 patients with early gastric cancer in a remnant stomach or a gastric tube.

Age, median (range), years

74 (53 – 84)

Sex, male/female, n

56 /3

Remnant stomach, n

48

Gastric tube, n

14

Reason for previous surgery, benign disease/cancer, n

Remnant stomach

17 /29

Gastric tube

 0 /13

Period between previous surgery and ESD, median (range), years

Remnant stomach for benign disease

30 (6 – 50)

Remnant stomach for gastric cancer

 6 (1 – 30)

Gastric tube for esophageal cancer

 4.5 (1 – 14)

Location of lesion in stomach, n

Upper

24

Middle

34

Lower

 4

Tumor diameter, median (range), mm

22 (5 – 59)

Pathological tumor depth

M

43

SM1

10

SM2

 9

Pathological ulceration

Present

19

Absent

43

Involving the anastomotic site or stump line

Yes

29

No

33

Indication for endoscopic resection

Standard

16

Expanded

29

Outside of criteria

17

ESD, endoscopic submucosal dissection; M, mucosal; SM1, submucosal invasion up to 500μm; SM2, submucosal invasion more than 500 μm.

Feasibility of ESD for EGC in a remnant stomach or gastric tube

[Table 2] shows the ESD treatment results for EGC in a remnant stomach or gastric tube and a whole stomach, according to the indications for endoscopic resection.

Table 2

Endoscopic submucosal dissection treatment results for early gastric cancer in a remnant stomach or gastric tube and a whole stomach, according to the indications for endoscopic resection.

ESD for remnant stomach or gastric tube

ESD for whole stomach

Total
(n = 62)

Standard
(n = 16)

Expanded
(n = 29)

Outside of criteria
(n = 17)

Standard
(n = 775)

Expanded
(n = 704)

Lesion size, median (IQR), mm

22
(14 – 33)

16[*]
(9 – 18)

22[*]
(13 – 30)

39[*]
(25 – 46)

11
(8 – 15)

28[*]
(22 – 37)

Presence of ulceration, n (%)

19 (31)

0

9 (31)

10 (26)

0

272 (39)

Involving anastomosis or stump line, n (%)

29 (47)

7 (44)

12 (41)

10 (59)

 – 

 – 

Operative time, median (IQR), minutes

66
(37 – 96)

40[*]
(22 – 55)

72[*]
(57 – 93)

90[*]
(78 – 133)

33
(19 – 45)

51
33 – 80)

En bloc resection, n (%)

59 (95)

16 (100)

27 (93)[#]

16 (94)

774 (99.8)

700 (99)[#]

En bloc with R0 resection, n (%)

53 (85)

15 (94)[#]

26 (90)

12 (71)

774 (98)[#]

665 (94)

Perforation, n (%)

11 (18)

1 (6)

7 (24)[#]

3 (18)

17 (2)

52 (7)[#]

Bleeding, n (%)

5 (8)

1 (6)

2 (7)

2 (12)

65 (8)

93 (13)

ESD, endoscopic submucosal dissection; IQR, interquartile range.

* Mann – Whitney’s U test: P < 0.05. All other comparisons non-significant.


# Chi-squared test: P < 0.05. All other comparisons non-significant.


The overall median operative time was 66 minutes (range 13 – 270 minutes, IQR 37.7−96 minutes). En bloc resection and en bloc with R0 resection were achieved in 59 (95 %) and 53 lesions (85 %), respectively. The median operative time for the standard indication group was significantly shorter (40 minutes) than that for the expanded criteria and outside the criteria groups (72 and 90 minutes, respectively; P < 0.05). Neither the en bloc resection rate nor en bloc with R0 resection rate was significantly different among the three indication groups.

Treatment results between lesions with or without involvement of the anastomotic site or stump line were compared ([Table 3]). The operative time was longer for lesions involving the anastomotic site or stump line (78 minutes) than that for those without involvement (60 minutes), but the difference was not significant. En bloc with R0 resection rate was apparently lower for lesions with involvement of the anastomotic site or stump line.

Table 3

Comparison of treatment results of endoscopic submucosal dissection in a remnant stomach or gastric tube between lesions with or without involvement of anastomotic site or stump line.

Involvement of anastomotic site or stump line

P value[*]

Yes
N = 29

No
n = 33

Operative time, median (IQR), minutes

78 (50 – 97)

60 (35 – 93)

n.s.

En bloc resection, n (%)

28 (97)

31 (94)

n.s.

En bloc with R0 resection, n (%)

23 (79)

30 (91)

n.s.

Perforation, n (%)

 8 (28)

 3 (9)

0.057

Bleeding, n (%)

 4 (14)

 1 (3)

n.s

Local recurrence, n (%)

 1 (4)

 1 (3)

n.s

IQR, interquartile range; n.s., not significant.

* Chi-squared test.



 

Comparison of treatment results between ESD in a remnant stomach or gastric tube and ESD in a whole stomach ([Table 2])

The operative time for standard (40 minutes) and expanded (72 minutes) criteria groups in a remnant stomach or gastric tube were longer than that for standard (33 minutes) and expanded (51 minutes) criteria groups in a whole stomach, however the differences were not significant.

The en bloc resection rate for lesions within the expanded criteria (93 %) and the en bloc with R0 resection rate for lesions within the standard criteria (94 %) for ESD in a remnant stomach or gastric tube were significantly lower (P < 0.05) than those of ESD in a whole stomach (99 % and 98 %, respectively).


 

Complications of ESD for EGC in a remnant stomach or gastric tube

As shown in [Table 2], the overall perforation rate was 18 % (11 /62) for ESD in a remnant stomach or gastric tube. Among the 11 cases, perforation occurred in lesions in the expanded criteria (seven cases) or outside the criteria group (three cases). The lesion involved the anastomotic site or stump line in eight cases, and ulcerative changes were observed in five lesions. The perforation rate for lesions within the expanded criteria was significantly higher (24 %) than that of lesions within the expanded criteria in a whole stomach (7 %) (P < 0.05). All of the perforations occurring during the ESD procedure were sealed immediately with endoclips. All patients were treated by fasting and with nasogastric suction and administration of antibiotics. No emergency surgery was required. The median hospitalized period for the perforation cases and non-perforation cases was 7 and 5 days, respectively (P = 0.04).

Postoperative bleeding occurred in five lesions (8 %). All hemorrhagic episodes were successfully treated by endoscopic clipping or coagulation, with no patient requiring a blood transfusion. Among the three lesion groups, the complication rates of perforation and bleeding were high in the expanded criteria and outside the criteria groups compared with the standard indication group but these differences were not significant.

The complication rate of perforation was as high as 28 % for lesions with involvement of the anastomotic site or stump line, which was higher than that of lesions without involvement (P = 0.057) ([Table 3]).


 

Recurrence and overall survival of ESD for EGC in a remnant stomach or a gastric tube

According to the pathology results, the ESD treatment was considered non-curative in 25 lesions (21 in the outside the criteria group, one with lymphovascular infiltration, and three lesions with positive surgical margins). Among these non-curative cases, six patients with a remnant stomach underwent additional surgery. The other patients refused surgery because of old age or coexisting diseases and were followed in an outpatient setting. Local recurrence occurred in two of these patients, at 14 and 15 months, respectively, during a median follow-up of 27 months after ESD. Both patients were considered to have undergone non-curative resections, but refused additional surgery. A second ESD was performed in one patient, and the other patient was treated by coagulation with a hot biopsy forceps and additional photodynamic therapy. The cumulative overall 3-year survival rate was 85 % ([Fig. 3]). No patient died from gastric cancer. Eight patients died from other malignancies, pulmonary infarction, and peritonitis. A metachronous EGC was found in two patients during the observation period, and these were treated by another ESD procedure.

Zoom
Fig. 3 Kaplan – Meier estimation of survival for patients who underwent endoscopic submucosal dissection in a remnant stomach or a gastric tube. The 3-year overall survival was 85 %. There were eight deaths from other causes during the observation period.

 

 

Discussion

Compared with surgery, endoscopic treatment including EMR is a minor invasive treatment and has been performed for the treatment of EGC in a remnant stomach or gastric tube. The reported en bloc resection rate of EMR performed for EGC in a remnant stomach or gastric tube is 14.3 % – 40.9 % [8] [10] [11]. ESD has the potential to achieve a higher en bloc resection rate than EMR, and several investigators have reported the application of ESD for EGC in a remnant stomach [9−12] or gastric tube [8] [13]. Among the case series of 13 to 40 patients, the en bloc resection rate of ESD in a remnant stomach has been reported as 95 %−100 % [9−12]. Among the case series of 8 – 10 patients of ESD in a gastric tube, the en bloc resection rate was reported as 88 %−90 % [8] [13]. The current study therefore represents the largest experience in this patient population, and showed a high en bloc resection rate of 95 %.

The en bloc resection rate of ESD for a remnant stomach or gastric tube was slightly lower than that of ESD in a whole stomach in our study. The en bloc resection rate was rather low for lesions involving the anastomotic site or stump line, or for lesions within the expanded or outside of criteria group, which included more lesions with ulcer findings. Previous studies of ESD in a whole stomach reported a marked decline in en bloc resection rates, from 92.9 % to 19.2 % in EGC with ulcer findings compared with EGC without ulcer findings [16]. In the current study, 31 % of the lesions had concomitant ulceration, and 47 % of the lesions involved the surgical suture line. The influence of fibrosis and deformity due to the previous surgery would be one reason to explain the decrease in the en bloc resection rate.

Although the operative time did not significantly differ among the same indication group (standard or expanded criteria), it was longer for ESD in a remnant stomach or gastric tube than in a whole stomach. The difference of lesion characteristics is that nearly half of the lesions included in the current study were those involving the anastomotic site or stump line. We believe that the anatomical deformity and limited working space due to the anastomosis or stump line are possible reasons for the time-consuming procedure.

A high rate of perforation (18 %) was identified in this study. Previous reports of perforation during ESD in a whole stomach have reported rates of 0 % to 12 % [17]. The presence of ulceration is reported to put the patient at a higher risk of perforation [18]. In the current study, except for one perforation case, all lesions had either ulceration or involved the anastomotic site or stump line. Once perforation occurs in a remnant stomach or a gastric tube, not only is pneumoperitoneum likely to occur but also pneumomediastinum, which may cause peritonitis or mediastinitis. Therefore, measures to avoid perforation should be taken, and rapid treatment is essential for minimizing complications. In the current study, all patients who experienced perforation were successfully treated by immediate closure with endoclips and subsequent nasogastric suction. Although no cases of delayed perforation occurred, we should always be aware of this possibility at low blood flow areas or fibrotic areas that have been subjected to too much electrocauterization [13].

In a large consecutive series, a 3-year overall survival rate of 98.4 % and a 3-year disease-specific survival rate of 100 % have been reported for patients who underwent ESD for an EGC in a whole stomach [19]. Although the numbers of patients in the current study are limited, results showed a 3-year overall survival rate of 85 %, and no deaths from gastric cancer were observed during the follow-up period.

In conclusion, this study represents the largest experience of ESD for EGC in a remnant stomach or gastric tube, with a high en bloc resection rate. However, this method is still a technically demanding treatment due to the high complication rate of perforation.


 

 

Competing interests: None.

Acknowledgments

This study was presented at the 18th United European Gastroenterology Week held in Barcelona, Spain, 23 – 27 October 2010.


Corresponding author

N. Kakushima, MD, PhD
Division of Endoscopy
Shizuoka Cancer Center
1007 Shimonagakubo
Nagaizumi
Suntogun
Shizuoka, 411-8777
Japan   
Phone: +81-55-9895222   
Fax: +81-55-9895692   


   Permissions and Reprints
Zoom
Fig. 1 Flow chart showing the patient inclusion in the study.
Zoom
Fig. 2 The endoscopic submucosal dissection procedure. a A reddish, slightly depressed lesion with a small nodule is observed in a narrow area of a gastric tube. b Chromoendoscopy with indigo carmine is performed to determine the border of the lesion. c Marking dots were placed around the lesion using an argon plasma coagulation probe. d, e Circumferential cutting was performed with an insulated-tip diathermic knife (IT Knife2). f The presence of staples was revealed during submucosal dissection. g The artificial ulcer after removing the lesion is carefully observed for residual vessels. h The lesion proved to be a mucosal cancer (blue lines, number 3 – 8), and was removed en bloc with negative margins.
Zoom
Zoom
Zoom
Zoom
Zoom
Zoom
Zoom
Zoom
Fig. 3 Kaplan – Meier estimation of survival for patients who underwent endoscopic submucosal dissection in a remnant stomach or a gastric tube. The 3-year overall survival was 85 %. There were eight deaths from other causes during the observation period.