Endoscopic submucosal dissection with or without snaring for colorectal neoplasms

ORIGINAL ARTICLE: Clinical Endoscopy

Endoscopic submucosal dissection with or without snaring for colorectal neoplasms Jeong-Sik Byeon, MD,1 Dong-Hoon Yang, MD,1 Kyung-Jo Kim, MD,1 Byong Duk Ye, MD,1 Seung-Jae Myung, MD,1 Suk-Kyun Yang, MD,1 Jin-Ho Kim, MD1 Seoul, Korea

Background: Despite a high en bloc resection rate, its technical difficulty and risk of complications limit the widespread use of colorectal endoscopic submucosal dissection (ESD). Objective: To analyze outcomes after colorectal ESD and ESD with snaring (ESD-S), a simplified modification of ESD. Design: A retrospective observational study. Setting: A single, tertiary-care, referral center. Patients and Intervention: ESD was performed on 163 lesions in 162 patients and ESD-S on 74 lesions in 71 patients. All lesions were nonpedunculated colorectal neoplasms of 15 mm or larger. Main Outcome Measurements: We analyzed procedure-related variables such as en bloc resection rate, procedure time, and complications. Histopathologic results were reviewed. Follow-up data were analyzed. Results: The en bloc resection rate was higher for ESD than for ESD-S (87% vs 64%, P ⬍ .01). The histologically complete resection rate was also higher for ESD. However, both rates for resection of lesions of ⬍20 mm were not different between ESD and ESD-S groups. The rates of perforation and bleeding were similar for both groups. Submucosal cancers were present in 21% and 31% of the ESD and ESD-S groups, respectively. One patient from each group developed a local adenoma recurrence. Limitations: Retrospective design. Conclusion: Both ESD and ESD-S were safe and effective for resection of colorectal neoplasms. The higher en bloc resection rate for ESD suggests that it should be the first option for resection of suspected superficial submucosal cancers in the colorectum. ESD-S can be a good alternative to ESD for en bloc resection of colorectal lesions of ⬍20 mm. (Gastrointest Endosc 2011;74:1075-83.)

Abbreviations: EMR-P, endoscopic mucosal resection after circumferential precutting; ESD, endoscopic submucosal dissection; ESD-S, endoscopic submucosal dissection with snaring. DISCLOSURE: All authors disclosed no financial relationships relevant to this publication. Copyright © 2011 by the American Society for Gastrointestinal Endoscopy 0016-5107/$36.00 doi:10.1016/j.gie.2011.03.1248 Received December 8, 2010. Accepted March 28, 2011. Current affiliations: Department of Gastroenterology (1), Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea. Reprint requests: Jeong-Sik Byeon, MD, Department of Gastroenterology, Asan Medical Center, 388-1 Pungnap-dong, Songpa-gu, Seoul 138-736, Korea. If you would like to chat with an author of this article, you may contact Dr Byeon at [email protected].

www.giejournal.org

Colorectal cancer is a major health problem in many regions, including Asia.1,2 Most colorectal cancers develop via colorectal adenomas, and therefore endoscopic resection of adenomas is an effective preventative strategy. However, unexpected interval invasive cancers can develop shortly after endoscopic resection of colorectal adenomas.3,4 Several mechanisms have been postulated for the development of such interval cancers, including incompleteness of endoscopic resection.4 Therefore, complete endoscopic resection is mandatory for the proper screening and surveillance of colorectal neoplasms. Hot snaring with or without submucosal saline solution injection is a commonly used endoscopic resection technique for colorectal neoplasms. Although easy and simple, snaring is not effective for achieving en bloc resection of large colorectal neoplasms, with the probability of pieceVolume 74, No. 5 : 2011 GASTROINTESTINAL ENDOSCOPY 1075

ESD for colorectal neoplasms

meal resection being as high as 49% to 100% for lesions larger than 2 to 3 cm.5-7 Endoscopic submucosal dissection (ESD) was originally developed to achieve en bloc resection of large gastric neoplasms.8 It has been used for large colorectal lesions, and the technique results in an en bloc resection rate of 73% to 99%.9-16 However, ESD has not been widely used for the removal of large colorectal neoplasms, possibly because, compared with other techniques, it is technically more difficult, takes longer, and has a higher risk of serious complications such as perforation. To address these issues, modifications to the ESD technique have been reported.17,18 One simplified modification to ESD is ESD with snaring (ESD-S). In the ESD-S procedure, standard ESD is initially undertaken; however, the final stage of the procedure involves the use of a snare to grasp and remove the undissected, narrowed submucosal tissue. We have previously reported on our initial experiences of using ESD-S for large colorectal lesions.19 The present study investigated the use of ESD and ESD-S in 233 patients with colorectal neoplasms. We analyzed procedure-related variables such as en bloc resection rate, procedure time and complications, histopathologic findings, and follow-up data.

PATIENTS AND METHODS Patients ESD and ESD-S were performed in 233 consecutive patients with colorectal neoplasms between November 2004 and June 2010. Both procedures were performed by a single endoscopist (J.S.B.), who had 4000 cases of diagnostic/therapeutic colonoscopy experience for 6 years and no prior experience with ESD/ESD-S before November 2004. He observed approximately 20 cases of colorectal ESD done in Japanese centers before this study. Trainees were not involved in any procedures. The indications for ESD or ESD-S were colorectal neoplasms that met the following criteria: (1) The lesion should be 15 mm or larger in size, (2) piecemeal resection would be predicted if endoscopic resection by conventional snaring techniques was used, and (3) deep submucosal cancer invasion was not suspected, based on white-light endoscopy and chromoscopy with indigo carmine. The endoscopic features indicating deep submucosal cancer invasion were expansion appearance, deeply depressed surfaces with irregular bottoms, folds converging toward the tumor, and a type VN pit pattern.20,21 A nonlifting sign also was a contraindication for ESD or ESD-S because of the risk of deep submucosal invasion.22 Pedunculated colorectal neoplasms with a definite stalk were excluded because they usually could be resected easily by using conventional snaring techniques. Informed consent was obtained from all patients. Data were obtained by reviewing med1076 GASTROINTESTINAL ENDOSCOPY Volume 74, No. 5 : 2011

Byeon et al

Take-Home Message ●

●

Both endoscopic submucosal dissection (ESD) and ESD with snaring may be effective and safe methods for the resection of colorectal neoplasms. ESD may be the most appropriate for en bloc resection of large, suspected superficial submucosal cancers in the colorectum, but ESD with snaring may be as effective as ESD for en bloc resection of relatively small colorectal neoplasms.

ical records. The study protocol was approved by the Institutional Ethics Committee of the Asan Medical Center.

ESD and ESD-S procedures For ESD, a single-channel endoscope (GIF-H260, GIFQ260J and/or CF-H260AL; Olympus Co, Tokyo, Japan) was used. A transparent hood was attached to the tip of the endoscope. After careful inspection of the lesion, a solution was injected submucosally around the lesion to lift it from the muscle layer. The solution was sodium hyaluronate (Shinpoong Co, Seoul, Korea) diluted with 1:100,000 epinephrine-normal saline solution (1:4) and mixed with a small amount of indigo carmine. The injection usually was performed on the anal side of the lesion first. However, it was performed on the oral side first if manipulation of the endoscope was easy with a retroflexed endoscope position. After the injection, the mucosa around the raised lesion was incised by using a fixed flexible snare knife (Kachu Technology Co, Seoul, Korea). The mucosa was incised at least 2 to 3 mm from the periphery of the lesion to create a tumor-free lateral resection margin. After a mucosal incision of approximately half of the lesion circumference, the lifted submucosal tissue beneath the lesion was dissected from the muscle layer by using the same knife. After submucosal dissection, the remainder of the unlifted lesion was lifted by an additional submucosal injection. The mucosal incision was then completed. Subsequent submucosal dissection was performed until the lesion was completely resected from the muscle layer (Fig. 1). After resection, visible vessels in the postESD ulcer were obliterated by using hemostatic forceps or argon plasma coagulation. Electrosurgical units (VIO300D; ERBE, Tuebingen, Germany) were used. Mucosal incisions and submucosal dissections were performed by using an Endocut Q current (effect 3, cut duration 2, cut interval 6) and a Forced coagulation current (effect 2, 40 W), respectively. The obliteration of visible vessels was achieved by using hemostatic forceps in Soft coagulation mode (effect 7, output 80 W). ESD-S involved initially performing ESD and then using snaring for the final resection step (Fig. 2). ESD was performed essentially as noted earlier, except that submucosal dissection was not performed until the lesion was www.giejournal.org

Byeon et al

ESD for colorectal neoplasms

Figure 1. Endoscopic submucosal dissection (ESD) of a colorectal neoplasm. A, A laterally spreading tumor approximately 4 ⫻ 4 cm was identified. B, We performed a submucosal injection and a mucosal incision. C, Submucosal dissection was done. D, Complete submucosal dissection was almost finished. E, A clear, post-ESD ulcer was created. F, The specimen was resected en bloc.

Figure 2. Endoscopic submucosal dissection (ESD) with snaring of a colorectal neoplasm. A, We identified a laterally spreading tumor of approximately 3 ⫻ 3 cm. B, A mucosal incision was performed. C, Submucosal dissection was done as per ESD. D, The final action was snaring (rather than the complete submucosal dissection performed in ESD). E, A clear ulcer was created. F, The specimen was resected en bloc.

completely separated from the muscle layer. Instead, submucosal dissection progressed circumferentially only to the degree that the undissected submucosal tissue was about 1 cm or less in diameter. The attached, narrowed submucosal tissue was then snared and resected by using an electric current. The choice of ESD or ESD-S was based on the endoscopist’s discretion. The main indications for ESD-S were (1) final snaring for faster resection in lesions for which snaring of the remnant undissected submucosal portion seemed easy, based on lesion location and lesion visibility on the video monitor and (2) final snaring to complete the resection in lesions for which location or accompanying www.giejournal.org

submucosal fibrosis made complete submucosal dissection difficult and could have resulted in a higher risk for perforation with ESD.23,24 Procedure time was defined as the time from the beginning of submucosal injection to complete resection of the lesion. Lesion size was endoscopically determined by using open biopsy forceps as a reference. Lesions were grossly classified as polypoid, nonpolypoid, or laterally spreading tumor.25 According to the Paris classification, a polypoid lesion included 0-Is lesions. Nonpolypoid lesions included 0-IIa, 0-IIb, 0-IIc, and combined ones such as 0-IIa ⫹ IIc and 0-IIc ⫹ IIa. A laterally spreading tumor was defined as a slightly elevated, nonprotruding lesion Volume 74, No. 5 : 2011 GASTROINTESTINAL ENDOSCOPY 1077

ESD for colorectal neoplasms

⬎20 mm in diameter.26 Laterally spreading tumor was subclassified into granular type and nongranular type. An en bloc resection was defined as a 1-piece resection. Immediate bleeding was defined as bleeding that developed during endoscopic resection, persisted despite endoscopic hemostasis, and eventually required intervention such as angiographic embolization and/or surgery. Delayed bleeding was defined as hematochezia and/or melena that occurred after the completion of endoscopic resection and required follow-up endoscopy for hemostasis. Perforation was identified based on endoscopic and/or radiological evidence.27

Histopathologic diagnosis Histopathologic diagnoses were made based on the Vienna classification.28 Histologically complete resection was defined as the absence of tumor cells at the resection margin of the specimen plus endoscopic en bloc resection. Endoscopic resection of cancer, based on previous studies,21,29 was considered curative if the following criteria were met: (1) well or moderately differentiated adenocarcinoma, (2) no lymphovascular invasion, (3) intramucosal cancer or superficial submucosal cancer, and (4) histologically complete resection, that is, no cancer cells at the resection margin. The definition of a superficial submucosal cancer was a depth of cancer invasion of sm 1 or ⬍1000 ␮m from the muscularis mucosa. The definition of a deep submucosal cancer was a depth of cancer invasion beyond that of a superficial submucosal cancer.

Follow-up after endoscopic resection Medical records were retrospectively reviewed to determine clinical outcomes. The first clinical follow-up was performed within 1 to 2 weeks after endoscopic resection. The occurrence of complications such as delayed bleeding was assessed. Further therapeutic plans also were made, based on the results of histopathologic examination, and additional surgical resection of the colon was recommended if a curative resection was not achieved in cancer cases. To evaluate local recurrence at the resection sites, all patients underwent periodic follow-up colonoscopies or sigmoidoscopies, the first of which was performed 2 to 6 months after endoscopic resection. Patients whose resected tumors were superficial submucosal cancers underwent periodic surveillance by using chest radiographs and abdominopelvic CT scans to monitor for any regional or distant metastases.

Statistical analysis Statistical analysis was performed by using PASW Statistics software (version 18.0, SPSS Inc, Chicago, Ill). Continuous data were compared by using a t test. Categorical data were analyzed by using the Pearson chi-square or Fisher exact tests. A logistic regression model was used for multivariate analysis of factors influencing en bloc resec1078 GASTROINTESTINAL ENDOSCOPY Volume 74, No. 5 : 2011

Byeon et al

tion. A P value of ⬍ .05 was considered to indicate significance.

RESULTS Baseline characteristics The study involved a total of 237 lesions in 233 patients. ESD was performed on 163 lesions in 162 patients, and ESD-S was performed on 74 lesions in 71 patients. There were 102 (63%) men in the ESD group and 44 (62%) men in the ESD-S group (P ⫽ .89). The mean (⫾ SD) ages were 61 ⫾ 10 years and 61 ⫾ 9 years in each group, respectively (P ⫽ .69). The characteristics of resected lesions are summarized in Table 1. The mean (⫾ SD) lesion size was greater in the ESD than in the ESD-S group (32.9 ⫾ 16.8 mm vs 23.8 ⫾ 9.3 mm; P ⬍ .01).

Procedure-related results In the ESD group, en bloc resection was achieved in 141 lesions (87%), and in the ESD-S group, en bloc resection was achieved in 47 lesions (64%) (P ⬍ .01). The procedure time was longer for ESD than for ESD-S (49 ⫾ 35 minutes vs 35 ⫾ 24 minutes; P ⬍ .01) (Table 1). The perforation rates were 7% for the ESD group and 3% for the ESD-S group, and these did not differ significantly (P ⫽ .24). All perforation cases were successfully managed without further surgery. The rates of immediate and delayed bleeding were similar for both techniques (Table 1). One immediate bleeding case was successfully managed by using angiographic embolization. All delayed bleeding cases were medically managed by using endoscopic hemostasis and transfusions. The en bloc resection rate, procedure time, and complication rates did not change significantly during the study period for both groups (data not shown).

Histopathologic results Two lesions in two ESD patients could not be resected endoscopically because of technical difficulties and immediate bleeding. Thus, histopathologic evaluation of endoscopically resected lesions was possible for 161 lesions (99%) of the 163 lesions in the ESD group. All 74 lesions of ESD-S patients were endoscopically resected and all were evaluated histopathologically. Adenoma was the most common lesion in both the ESD and ESD-S groups (Table 2). The histologically complete resection rate was higher in the ESD group than in the ESD-S group (75% vs 57%; P ⫽ .01). The proportions of tumor-negative lateral and deep resection margins were also higher in the ESD group (Table 2). Twenty-two of 27 patients with deep submucosal cancer underwent further surgical colonic resection with lymph node dissection. One patient underwent radiation therapy after endoscopic resection of a deep submucosal cancer in the rectum. The other 4 patients refused any additional therapy. www.giejournal.org

Byeon et al

ESD for colorectal neoplasms

TABLE 1. Characteristics of resected lesions and procedure-related variables

Characteristic Lesion size, mean (SD), mm

TABLE 2. Histopathologic analyses of resected specimens

ESD (n ⴝ 163)

ESD-S (n ⴝ 74)

P value

Finding, no. (%)

32.9 (⫾ 16.8)

23.8 (⫾ 9.3)

⬍ .01

Histopathology

Lesion gross type, no. (%)

.01

LST-G

66 (41)

22 (30)

LST-NG

49 (30)

14 (19)

Polypoid

38 (23)

31 (42)

Nonpolypoid

10 (6)

7 (10)

Lesion location, no. (%)

.67

Rectum

86 (53)

38 (51)

SC

25 (15)

15 (20)

DC

6 (4)

4 (5)

TC

19 (12)

5 (7)

AC-cecum

27 (17)

12 (16)

Procedure time, mean (SD), min

49 (⫾ 35)

35 (⫾ 24)

En bloc resection, no. (%)

141 (87)

47 (64)

ESD-S (n ⴝ 74)

Adenoma

86 (53)

39 (53)

Intramucosal cancer

41 (26)

12 (16)

Superficial submucosal cancer

14 (9)

16 (22)

Deep submucosal cancer

20 (12)

7 (10)

120 (75)

42 (57)

Histologically complete resection

⬍ .01

Tumor negative

124 (77)

43 (58)

Tumor positive

17 (11)

4 (5)

Unknown*

20 (12)

27 (37) ⬍.01

Tumor negative

136 (85)

45 (61)

Tumor positive

5 (3)

2 (3)

20 (12)

27 (37)

Unknown*

.01 ⬍.01

Deep resection margin ⬍ .01

P value .03

Lateral resection margin

ESD, Endoscopic submucosal dissection; ESD-S, ESD with snaring. *Cases in which the histological involvement of tumor at the resection margin could not be properly evaluated because the tumor was resected in a piecemeal manner.

Complications, no. (%) Immediate bleeding

1 (1)

0 (0)

1.00

Delayed bleeding

1 (1)

2 (3)

.23

Perforation

12 (7)

2 (3)

.24

ESD, Endoscopic submucosal dissection; ESD-S, ESD with snaring; SD, standard deviation; LST-G, laterally spreading tumor– granular type; LST-NG, laterally spreading tumor–nongranular type; SC, sigmoid colon; DC, descending colon; TC, transverse colon; AC, ascending colon.

Curative resection was achieved in 27 of 30 patients with superficial submucosal cancer. Two of the 27 patients with curative resection underwent additional surgery on patient request. Three patients in whom curative resection was not achieved because of lymphovascular invasion refused additional surgery.

Subanalysis according to lesion size and gross type Outcome analyses according to lesion size are summarized in Table 3. In colorectal neoplasms ⱖ20 mm, ESD was superior to ESD-S in terms of higher en bloc resection and histologically complete resection rates. However, in www.giejournal.org

ESD (n ⴝ 161)

lesions ⬍20 mm, ESD-S was similar to ESD in terms of en bloc resection, histologically complete resection, and complication rates. The en bloc resection rate was analyzed according to the gross type of lesions in both ESD and ESD-S groups (Table 4). The en bloc resection rate was not different between laterally spreading tumor– granular type and laterally spreading tumor–nongranular type in both ESD and ESD-S groups.

Factors influencing en bloc resection We performed multivariate analysis to assess factors influencing en bloc resection. ESD was superior to ESD-S independent of other factors. Smaller lesion size was also an independent predictive factor for en bloc resection. Other variables such as the gross type, location, and histopathology of lesions were not independent factors influencing en bloc resection (Table 5).

Follow-up data Follow-up endoscopies were performed for 148 lesions (92 ESD and 56 ESD-S lesions). The median numVolume 74, No. 5 : 2011 GASTROINTESTINAL ENDOSCOPY 1079

ESD for colorectal neoplasms

Byeon et al

TABLE 3. Outcomes between ESD and ESD-S groups according to lesion size Lesion <20 mm

Lesion >20 mm

ESD (n ⴝ 22)

ESD-S (n ⴝ 27)

P value

ESD (n ⴝ 141)

ESD-S (n ⴝ 47)

P value

31 (⫾ 13)

25 (⫾ 18)

.23

52 (⫾ 37)

40 (⫾ 25)

.05

En bloc resection, no. (%)

21 (96)

24 (89)

.62

120 (85)

23 (49)

⬍ .01

Histologically complete resection, no. (%)

19 (86)

22 (82)

.72

101 (73)

20 (43)

⬍ .01

Immediate bleeding

0 (0)

0 (0)

1.00

1 (1)

0 (0)

1.00

Delayed bleeding

0 (0)

0 (0)

1.00

1 (1)

2 (4)

.16

Perforation

3 (14)

0 (0)

.08

9 (6)

2 (4)

.73

Outcome Procedure time, mean (SD), min

Complications, no. (%)

ESD, Endoscopic submucosal dissection; ESD-S, ESD with snaring; SD, standard deviation.

TABLE 4. En bloc resection rates according to gross type of lesion Lesion, no (%)

LST-G

LST-NG

Polypoid

Nonpolypoid

P value

ESD (n ⫽ 163)

60/66 (91)

45/49 (92)

26/38 (68)

10/10 (100)

.01

ESD-S (n ⫽ 74)

13/22 (59)

8/14 (57)

20/31 (65)

6/7 (86)

.62

LST-G, Laterally spreading tumor– granular type; LST-NG, laterally spreading tumor–nongranular type; ESD, endoscopic submucosal dissection; ESD-S, ESD with snaring.

ber and period of follow-up endoscopies were 2 (range 1-6) for 13 months (range 2-52 months) and 2 (range 1-4) for 20 months (range 2-59 months) in each group, respectively (P ⬍ .01). One (1%) of the 92 ESD lesions in patients who underwent follow-up endoscopies showed local adenoma recurrence at the primary ESD site. In that lesion, the lateral resection margin of the initial ESD specimen showed adenoma involvement. Similarly, 1 (2%) of the 56 ESD-S lesions developed local adenoma recurrence at the primary ESD-S site, and the lesion was initially removed piecemeal. Follow-up chest radiographs and abdominopelvic CT scans were performed in 3 patients who refused additional surgery after ESD or ESD-S for deep submucosal cancers and in 18 patients with superficial submucosal cancers. The median number and period of follow-up chest radiographs and abdominopelvic CT scans was 2 (range 1-7) for 20 months (range 6-59 months) and 3 (range 1-8) for 22 months (range 6-57 months), respectively. No distant metastasis was detected. Follow-up data are shown in Figure 3.

DISCUSSION The present study found that both ESD and ESD-S were safe and effective treatment approaches for colorectal neoplasms. ESD-S is a method similar to endoscopic mucosal resection after circumferential precutting (EMR-P) that has 1080 GASTROINTESTINAL ENDOSCOPY Volume 74, No. 5 : 2011

been used for the resection of gastric lesions.30,31 However, ESD-S uses submucosal dissection between mucosal precutting and final snaring, which is not performed in EMR-P. We postulated that the process of submucosal dissection might improve the probability of en bloc resection. When submucosal dissection is sufficiently done, that is, until the undissected submucosal tissue was about 1 cm or less in diameter in this study, the snare can be reliably placed between the dissected portion of the lesion and the corresponding submucosal bed. The snare can then be narrowed without slipping out of the dissected space, and the whole lesion can be successfully grasped. We believe that this theoretical advantage resulted in the high ESD-S en bloc resection rate (64%), which appeared to be higher than the 0% to 51% rate reported after conventional snaring for large colorectal lesions.5-7 However, the en bloc resection rate for ESD-S was lower than that for ESD (87%). It is unclear why the en bloc resection rate was lower for ESD-S than for ESD in the present study. One reason may be that it is sometimes difficult to be certain a snare has completely grasped the oral side of a lesion, especially for a large lesion. In addition, the final thickness of undissected submucosal tissue before snaring in ESD-S was assessed only by endoscopic view during the procedure. Therefore, the thickness of undissected submucosal tissue at the time of snaring might be over 1 cm in some cases. Insufficient submucosal dissection can result in slipping of the snare, resulting in a failed en bloc resection. Studies www.giejournal.org

Byeon et al

ESD for colorectal neoplasms

TABLE 5. Multivariate analysis for factors influencing en bloc resection Factor

OR (95% CI)

ESD-S ESD Increasing size*

Reference 6.48 (2.87, 14.60) 0.97 (0.95, 0.99)

Location Rectum SC-cecum

Reference 0.83 (0.38, 1.81) .06 Reference

LST-NG

0.62 (0.22, 1.78)

Polypoid

0.38 (0.16, 0.93)

Nonpolypoid

3.05 (0.30, 30.51)

Histopathology Adenoma

.01 .83

Gross type LST-G

P value ⬍.01

Resection method

.40 Reference

Intramucosal cancer

0.65 (0.26, 1.61)

Superficial submucosal cancer

1.71 (0.52, 5.57)

Deep submucosal cancer

0.56 (0.17, 1.85)

OR, Odds ratio; CI, confidence interval; ESD-S, endoscopic submucosal dissection with snaring; ESD, endoscopic submucosal dissection; SC, sigmoid colon; LST-G, laterally spreading tumor– granular type; LST-NG, laterally spreading tumor–nongranular type. *Size was regarded as a continuous variable.

aimed at defining the optimal thickness of undissected submucosal tissue and the objective method to assess the thickness of the undissected portion before snaring may improve the ESD-S en bloc resection rate. Because the size of lesions was different between ESD and ESD-S groups, we performed subanalyses according to lesion size. Although the en bloc resection and histologically complete resection rates were lower for ESD-S in overall lesions, those for ESD-S in small lesions of ⬍20 mm were 89% and 82%, respectively, which were statistically the same as those for ESD. This finding is the same as that of the previous study analyzing ESD-S of gastric lesions.23 Complication rates also were not different between ESD and ESD-S groups. Because ESD-S is obviously simpler and easier than ESD, we suggest that ESD-S can be a good option in colorectal lesions of 15 to 19 mm in size for which piecemeal resection would be predicted if endoscopic resection by conventional snaring techniques were used. A comprehensive histopathologic evaluation of an endoscopically resected specimen increases the likelihood of www.giejournal.org

cure in submucosal colorectal cancer patients.32 Such comprehensive evaluation requires an en bloc specimen. In addition, the risk of local recurrence at the primary resection site is lower after en bloc resection than after piecemeal resection.5-7 Considering these, along with the higher en bloc resection rate for ESD than ESD-S in the present study, we suggest that ESD may be the most appropriate treatment for suspected submucosal colorectal cancers of ⱖ20 mm. This suggestion can be supported by the colorectal ESD indications reported by a Japanese group, which included carcinomas ⬎20 mm with submucosal infiltration.33 We suggest a management strategy for nonpedunculated colorectal polyps in Figure 4, based on our study. This strategy was devised mainly in terms of en bloc resection. Comparison of long-term outcome and cost effectiveness between ESD, ESD-S, and conventional snaring techniques, including piecemeal resection, should be performed for the validation of this strategy. In the present study, perforation occurred in 7% and 3% of patients after ESD and ESD-S, respectively, and the difference was not statistically significant. One case developed bleeding necessitating angiographic embolization. The present complication rates after ESD and ESD-S are similar to rates reported by previous ESD studies.9-16 Because all complications were successfully managed without further surgery in our study, we believe that both ESD and ESD-S may be safe procedures. In the present study, only one (1%) ESD lesion and one (2%) ESD-S lesion developed adenoma recurrence at the resection sites. Conventional snaring techniques such as piecemeal resection are associated with a local recurrence rate of 3% to 22%.5-7 Therefore, it appears that ESD and ESD-S are superior to conventional techniques in terms of risk of local recurrence. Most of the present cases with a positive resection margin or piecemeal resection did not develop recurrence (Fig. 3). This is probably because the largest portion of each lesion was resected in one or two pieces by submucosal dissection even if en bloc resection failed, resulting in a low rate of multiple intralesional cuts. In contrast, conventional endoscopic piecemeal resection often results in multiple intralesional cuts, which are associated with a high risk of remnant lesion between intralesional cuts and resultant local recurrence. In addition, in some cases, the fragile lateral mucosal edge of a resected specimen can be partly torn off during specimen retrieval, resulting in a false-positive resection margin.9 The current study had some limitations. First, the study was not randomized in design, and the selection of the resection method was not made before the procedure. Therefore, there may be a selection bias. Also, the two groups differed in terms of number and mean lesion size, which can compromise comparative analysis. However, we performed multivariate analysis for factors influencing en bloc resection so as to overcome these differences, Volume 74, No. 5 : 2011 GASTROINTESTINAL ENDOSCOPY 1081

ESD for colorectal neoplasms

Byeon et al

Figure 3. Follow-up results after ESD and ESD with snaring for colorectal neoplasms. ESD, endoscopic submucosal dissection; ESD-S, ESD with snaring.

Figure 4. A strategic approach for endoscopic resection of nonpedunculated colorectal neoplasms.

which showed superiority of ESD to ESD-S independent of other factors such as size and location of lesions. Second, some ESD-S was performed for the lesions in difficult locations or having submucosal fibrosis, which might result in a lower en bloc resection rate in the ESD-S group. Third, the endoscopist’s skill change over the study period was not considered. Finally, the follow-up period was not long, and the proportion of patients who were lost to follow-up was not small. These made reliable analyses of long-term outcomes difficult. However, the strengths of the study are that it enrolled a relatively large number of patients undergoing both ESD and ESD-S and that outcome data regarding the latter technique are scarce for colorectal lesions. In conclusion, ESD and ESD-S were both found to be safe and effective for resection of colorectal neoplasms. The higher en bloc resection rate for ESD suggests that it should be the first option for resection of suspected superficial submucosal cancers in the colorectum. ESD-S can be a good alternative to ESD for en bloc resection of colorectal lesions of ⬍20 mm. Large scale, prospective, randomized studies are warranted to confirm these suggestions. REFERENCES 1. Jemal A, Siegel R, Xu J, et al. Cancer statistics, 2010. CA Cancer J Clin 2010;60:277-300.

1082 GASTROINTESTINAL ENDOSCOPY Volume 74, No. 5 : 2011

2. Sung JJ, Lau JY, Goh KL, et al; Asia Pacific Working Group on Colorectal Cancer. Increasing incidence of colorectal cancer in Asia: implications for screening. Lancet Oncol 2005;6:871-6. 3. Martínez ME, Baron JA, Lieberman DA, et al. A pooled analysis of advanced colorectal neoplasia diagnoses after colonoscopic polypectomy. Gastroenterology 2009;136:832-41. 4. Farrar WD, Sawhney MS, Nelson DB, et al. Colorectal cancers found after a complete colonoscopy. Clin Gastroenterol Hepatol 2006;4:1259-64. 5. Bergmann U, Beger HG. Endoscopic mucosal resection for advanced non-polypoid colorectal adenoma and early stage carcinoma. Surg Endosc 2003;17:475-9. 6. Hurlstone DP, Sanders DS, Cross SS, et al. Colonoscopic resection of lateral spreading tumours: a prospective analysis of endoscopic mucosal resection. Gut 2004;53:1334-9. 7. Conio M, Repici A, Demarquay JF, et al. EMR of large sessile colorectal polyps. Gastrointest Endosc 2004;60:234-41. 8. Gotoda T. A large endoscopic resection by endoscopic submucosal dissection procedure for early gastric cancer. Clin Gastroenterol Hepatol 2005;3:S71-3. 9. Fujishiro M, Yahagi N, Nakamura M, et al. Endoscopic submucosal dissection for rectal epithelial neoplasia. Endoscopy 2006;38:493-7. 10. Onozato Y, Kakizaki S, Ishihara H, et al. Endoscopic submucosal dissection for rectal tumors. Endoscopy 2007;39:423-7. 11. Hurlstone DP, Atkinson R, Sanders DS, et al. Achieving R0 resection in the colorectum using endoscopic submucosal dissection. Br J Surg 2007;94:1536-42. 12. Tanaka S, Oka S, Kaneko I, et al. Endoscopic submucosal dissection for colorectal neoplasia: possibility of standardization. Gastrointest Endosc 2007;66:100-7. 13. Saito Y, Uraoka T, Matsuda T, et al. Endoscopic treatment of large superficial colorectal tumors: a case series of 200 endoscopic submucosal dissections (with video). Gastrointest Endosc 2007;66:966-73.

www.giejournal.org

Byeon et al

ESD for colorectal neoplasms

14. Tamegai Y, Saito Y, Masaki N, et al. Endoscopic submucosal dissection: a safe technique for colorectal tumors. Endoscopy 2007;39:418-22. 15. Zhou PH, Yao LQ, Qin XY. Endoscopic submucosal dissection for colorectal epithelial neoplasm. Surg Endosc 2009;23:1546-51. 16. Saito Y, Fukuzawa M, Matsuda T, et al. Clinical outcome of endoscopic submucosal dissection versus endoscopic mucosal resection of large colorectal tumors as determined by curative resection. Surg Endosc 2010;24:343-52. 17. Uraoka T, Kato J, Ishikawa S, et al. Thin endoscope-assisted endoscopic submucosal dissection for large colorectal tumors (with videos). Gastrointest Endosc 2007;66:836-9. 18. Sakamoto N, Osada T, Shibuya T, et al. Endoscopic submucosal dissection of large colorectal tumors by using a novel spring-action S-O clip for traction (with video). Gastrointest Endosc 2009;69:1370-4. 19. Byeon JS, Jo JY, Choi KD, et al. Combination of endoscopic submucosal dissection and snare resection in colorectal lesions. J Gastroenterol Hepatol 2005;20:A246-7. 20. Saitoh Y, Obara T, Watari J, et al. Invasion depth diagnosis of depressed type early colorectal cancers by combined use of videoendoscopy and chromoendoscopy. Gastrointest Endosc 1998;48:362-70. 21. Kudo S, Kashida H, Tamura T, et al. Colonoscopic diagnosis and management of nonpolypoid early colorectal cancer. World J Surg 2000;24:1081-90. 22. Ishiguro A, Uno Y, Ishiguro Y, et al. Correlation of lifting versus nonlifting and microscopic depth of invasion in early colorectal cancer. Gastrointest Endosc 1999;50:329-33. 23. Goto O, Fujishiro M, Kodashima S, et al. Feasibility of electrocautery snaring as the final step of endoscopic submucosal dissection for stomach epithelial neoplasms. Dig Liver Dis 2009;41:26-30. 24. Nishiyama H, Isomoto H, Yamaguchi N, et al. Endoscopic submucosal dissection for colorectal epithelial neoplasms. Dis Colon Rectum 2010; 53:161-8.

25. The Paris endoscopic classification of superficial neoplastic lesions: esophagus, stomach, and colon: November 30 to December 1, 2002. Gastrointest Endosc 2003;58:S3-43. 26. Endoscopic Classification Review Group. Update on the paris classification of superficial neoplastic lesions in the digestive tract. Endoscopy 2005;37:570-8. 27. Yang DH, Byeon JS, Lee KH, et al. Is endoscopic closure with clips effective for both diagnostic and therapeutic colonoscopy-associated bowel perforation? Surg Endosc 2010;24:1177-85. 28. Schlemper RJ, Riddell RH, Kato Y, et al. The Vienna classification of gastrointestinal epithelial neoplasia. Gut 2000;47:251-5. 29. Kitajima K, Fujimori T, Fujii S, et al. Correlations between lymph node metastasis and depth of submucosal invasion in submucosal invasive colorectal carcinoma: a Japanese collaborative study. J Gastroenterol 2004;39:534-43. 30. Min BH, Lee JH, Kim JJ, et al. Clinical outcomes of endoscopic submucosal dissection (ESD) for treating early gastric cancer: comparison with endoscopic mucosal resection after circumferential precutting (EMR-P). Dig Liver Dis 2009;41:201-9. 31. Choi IJ, Kim CG, Chang HJ, et al. The learning curve for EMR with circumferential mucosal incision in treating intramucosal gastric neoplasm. Gastrointest Endosc 2005;62:860-5. 32. Bond JH. Polyp guideline: diagnosis, treatment, and surveillance for patients with colorectal polyps. Practice Parameters Committee of the American College of Gastroenterology. Am J Gastroenterol 2000;95: 3053-63. 33. Tanaka S, Oka S, Chayama K. Colorectal endoscopic submucosal dissection: present status and future perspective, including its differentiation from endoscopic mucosal resection. J Gastroenterol 2008;43: 641-51.

Availability of Journal back issues As a service to our subscribers, copies of back issues of Gastrointestinal Endoscopy for the preceding 5 years are maintained and are available for purchase from Elsevier until inventory is depleted. Please write to Elsevier Inc., Subscription Customer Service, 3251 Riverport Lane, Maryland Heights, MO 63043 or call 800-654-2452 or 314-447-8871 for information on availability of particular issues and prices.

www.giejournal.org

Volume 74, No. 5 : 2011 GASTROINTESTINAL ENDOSCOPY 1083