Outcomes of Endoscopic Submucosal Dissection for Colorectal Epithelial Neoplasms in 200 Consecutive Cases

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2007;5: ORIGINAL ARTICLES Outcomes of Endoscopic Submucosal Dissection for Colorectal Epithelial Neoplasms in 200 Consecutive Cases MITSUHIRO FUJISHIRO,* NAOHISA YAHAGI,* NAOMI KAKUSHIMA,* SHINYA KODASHIMA,* YOSUKE MURAKI,* SATOSHI ONO,* NOBUTAKE YAMAMICHI,* AYAKO TATEISHI,* MASASHI OKA, KEIJI OGURA,* TAKAO KAWABE,* MASAO ICHINOSE, and MASAO OMATA* *Department of Gastroenterology, Graduate School of Medicine, University of Tokyo, Tokyo; and Second Department of Internal Medicine, Wakayama Medical University, Wakayama, Japan See CME exam on page 645. Background & Aims: The clinical outcomes for endoscopic submucosal dissection (ESD), a novel endoluminal surgery for gastrointestinal neoplasm in the colorectum, are reported. Methods: ESD was performed on 186 consecutive patients with 200 colorectal epithelial neoplasms who had preoperative diagnoses of mucosal or slight submucosally invasive neoplasms. In addition, these could be of large size, with submucosal fibrosis, or located on an intestinal fold. The therapeutic efficacy and safety were assessed. Results: The targeted lesions consisted of 102 adenomas, 72 noninvasive carcinomas, and 26 invasive carcinomas. Seven lesions (3.5%) were histologically considered to be at substantial risk for nodal metastasis after ESD. The rate of en bloc resection was 91.5% (183/200), and en bloc resection with tumor-free lateral/basal margins (R0 resection) was 70.5% (141/200). Two lesions (1%) required emergency colonoscopies as a result of hematochezia after ESD. Eleven (5.5%) immediate perforations that occurred during ESD were successfully managed conservatively, but 1 (0.5%) delayed perforation required laparotomy. Two multiple-piece resections of 111 tumors (1.8%), which were successfully followed by colonoscopy (median follow-up, 18 months; range, months), were found as locally recurrent tumors 2 and 21 months after ESD. No lymph node or distant metastasis was detected in 77 patients with noninvasive or invasive carcinoma (median follow-up, 24 months; range, 6 74 months). Conclusions: ESD is applicable in the colorectum with promising results. However, when considering the risks and benefits, piecemeal endoscopic resection or colorectal resection might be more appropriate for some subgroups of large flat neoplasms or those with submucosal fibrosis. The importance of small, flat, and depressed colorectal epithelial neoplasms was stressed after the recognition of de novo carcinoma caused by a high probability of harboring foci of invasive carcinoma. 1,2 However, early detection of these small tumors allows them to be endoscopically resected in an en bloc fashion by using the inject and cut endoscopic mucosal resection (EMR) technique. 3 On the contrary, the therapeutic approach for large, flat colorectal epithelial neoplasms, ie, laterally spreading tumors (LSTs), is still under development. These tumors are commonly surgically resected with colorectal resection or endoscopically resected in multiple pieces. 4,5 The latter is not preferred from an oncologic standpoint because of insufficient histologic assessment and low curability. Furthermore, repeated endoscopic resection is not suitable for some non-lifting recurrent tumors as a result of the high risk of perforation. A newly developed endoluminal surgery, endoscopic submucosal dissection (ESD), was originally developed for en bloc resection of large or ulcerative gastric epithelial neoplasms. 6 With this technique, a large colorectal neoplasm or a non-lifting intraepithelial neoplasm with submucosal fibrosis caused by previous endoscopic treatment or biopsy might be endoscopically resected in an en bloc fashion. We previously reported promising efficacy, albeit substantial risk, for applying ESD to the rectum. 7 However, the outcomes in the whole colorectum have not been previously reported. In this study, we assessed the outcomes of ESD for colorectal epithelial neoplasms with our own technique. Patients and Methods Between July 2000 March 2006, 186 consecutive patients, who had a total of 200 colorectal epithelial neoplasms, gave written informed consent to be treated with ESD at our hospital. Surgery was performed by 2 ESD specialists who were highly experienced in performing ESD in the stomach. Lesions with an indication for ESD were determined by endoscopic features by using chromoendoscopy with or without magnifying endoscopy. In addition, endoscopic ultrasonography was Abbreviations used in this paper: EMR, endoscopic mucosal resection; ESD, endoscopic submucosal dissection; LST, laterally spreading tumor; LST-G-H, laterally spreading tumor, granular and homogenous type; LST-G-M, laterally spreading tumor, granular and nodular mixed type; LST-NG-F, laterally spreading tumor, non-granular and flat elevated type; LST-NG-PD, laterally spreading tumor, non-granular and pseudodepressed type; SCAR, submucosal fibrosis showing non-lifting signs caused by previous endoscopic treatment or biopsy; SM, submucosal invasive carcinoma by the AGA Institute /07/$32.00 doi: /j.cgh

2 June 2007 COLORECTAL ENDOSCOPIC SUBMUCOSAL DISSECTION 679 performed for lesions with a high probability of submucosal invasion. One of the following criteria was preoperatively predicted in the subjects of ESD: (1) intraepithelial neoplasms 2 cm in size or on the colorectal fold that might result in multipiece resection by conventional EMR; (2) intraepithelial neoplasms with submucosal fibrosis caused by previous endoscopic treatment or biopsy that might show non-lifting signs; or (3) invasive carcinomas with a slight submucosal penetration (submucosal invasive carcinoma [SM]1, or 1000 m below the muscularis mucosae). Some patients with carcinoid tumors were resected by ESD during this period, but these patients were excluded from the analysis because of the different biologic behavior of these lesions. Some invasive carcinomas that were treated by ESD as a result of the patient s desire or in a palliative fashion under a preoperative diagnosis of massive submucosal penetration (SM2, or 1000 m below the muscularis mucosae) were also excluded because of the extra-indication of curative ESD. Endoscopic characteristics of the tumors were classified according to our modification of the Paris endoscopic classification. 8 Macroscopic types of indicated tumors were divided into protruding large tumor (type 0-I); 4 subtypes of LST according to Kudo s classification (granular and homogenous type [LST- G-H], granular and nodular mixed type [LST-G-M], non-granular and flat elevated type [LST-NG-F], and non-granular and pseudodepressed type [LST-NG-PD]) 9 ; or intraepithelial tumor with submucosal fibrosis showing non-lifting signs caused by previous endoscopic treatment or biopsy (SCAR) (Figure 1). Adequate cleansing of the whole colorectum was conducted before performing ESD. The patients were restricted to a lowfiber diet the day before ESD, and 10 ml of 0.75% sodium picosulfate solution (Laxoberon; Teijin Pharma Co, Ltd, Tokyo, Japan) was used the night before the ESD procedure. Early on the morning of the ESD procedure, 10 mg of mosapride citrate (Gasmotin; Dainippon Sumitomo Pharma Co, Ltd, Osaka, Japan) and 2 L of an isotonic polyethylene glycol electrolyte solution (Niflec; Ajinomoto Pharma Co, Ltd, Tokyo, Japan) were used to achieve good bowel preparation. The details of the ESD procedure have been reported previously (Figure 2). 10,11 ESD was principally carried out by using a single-channel upper gastrointestinal endoscope with a waterjet system (GIF-Q260J, Olympus or EG-2931, Pentax, Tokyo, Japan). Alternatively, a slim variable stiffness colonoscope without a water-jet system (PCF-Q240AI, Olympus) was used for cases of deep proximal colon that were difficult to reach with the upper gastrointestinal endoscopes. Deep insertion was applied as far as possible to wash residue out of the colorectum, even if the target lesion was located in the distal colon or the rectum. The electrosurgical unit used was ICC 200 (from July 2000 March 2005) or VIO 300D (from April 2005). Both were products of ERBE Elektromedizin, Tübingen, Germany, which used a special cutting current, the ENDOCUT mode. A mixture of a 1% 1900 kd hyaluronic acid preparation (Suvenyl; Chugai Pharmaceutical Co, Tokyo, Japan) plus normal saline (from July 2000 October 2003) or 10% glycerin plus 5% fructose and 0.9% saline preparation (Glyceol; Chugai Pharmaceutical Co) (from November 2003) was used as a submucosal injection solution. The mixing solution with hyaluronic acid preparation was changed because of new knowledge of submucosal injection solutions. 12,13 The mixing ratio of the former and the latter was also changed from 1:3 (from July 2000 March 2004) to 1:7 (from April 2004) because of technical advances. The principal electrosurgical knife was a tip of electrosurgical snare (thin type) (SD-7p-1; Olympus) (from July 2000 October 2002) or Flexknife (KD-630L; Olympus) (from November 2002). 14,15 In difficult dissecting situations, Hookknife (KD-620LR; Olympus) 16 was used in combination with Flexknife. To control bleeding, hemostatic forceps (SDB2422, Pentax or FD-410LR, Olympus) were also used. 17 Figure1. Endoscopic features of colorectal neoplasms for which ESD was indicated. (A) Protruding large tumor (type 0-I); (B) Laterally spreading tumor (LST)-G-H; (C) LST-G-M; (D) LST-NG-F; (E) LST-NG-PD; (F) tumor with SCAR.

3 680 FUJISHIRO ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 5, No. 6 Figure 2. ESD of colorectal neoplasms. (A) Chromoendoscopic view with indigo carmine dye showing demarcation of the margin of an LST-NG-F, intramucosal adenocarcinoma, 2.5 cm in size, located in the ascending colon. (B) Submucosal injection at the oral margin of the lesion with the retroflexed position of the endoscope. (C) Initial mucosal incision at the oral margin of the lesion with the retroflexed position of the endoscope. (D) Submucosal injection at the anal margin with the straight position of the endoscope. (E) Mucosal incision at the anal margin and extension of the incision in a circumferential manner around the lesion with the straight position of the endoscope. (F) Repetition of submucosal injections from the exposed submucosal layer and dissection of the submucosal connective tissue until the lesion detached. (G) Artificial ulcer after removal. (H) Complete resection of the lesion in one piece. Immediately after ESD, patients were permitted small amounts of water in bed. On the following day, if the patient s symptoms, laboratory findings, and abdominal x-ray were unremarkable, a light meal was permitted, and the patient was discharged within 1 week. If complications occurred, the schedules were changed according to the individual patient s condition. Histologic classification was performed microscopically according to the revised Vienna classification of gastrointestinal epithelial neoplasm. 18,19 Because submucosal massive invasion (SM2 or deeper, 1000 m below the muscularis mucosae), existence of poorly differentiated adenocarcinoma, and/or vessel infiltration are regarded as high risks for positive lymph nodes, additional surgical intervention was strongly recommended in these cases. 8,20 Extension of tumor cells to the resected margin was evaluated as complete (R0) resection with the lateral and basal resection margins free of tumor (en bloc resection is essential), incomplete (R1) resection when the tumor extended into the lateral or basal margins, or not evaluable (Rx) when the margins were not evaluable as a result of the artificial effects of coagulation necrosis or multi-piece resection. Follow-up colonoscopy with chromoendoscopy was usually performed about 2 months after ESD to confirm healing of the post-procedure ulcers and to exclude the presence of residual tumors. This was followed by annual endoscopic examinations to check for local recurrence and/or second primary tumors. In case of noninvasive or invasive carcinoma, the existence of distant or lymph node metastases was indefinitely evaluated through medical checkups, including computed tomography, by the doctor in charge. Statistical analysis between groups was performed by using the 2 test with or without Yates continuity correction, and a P value.05 was considered a significant difference. Results Table 1 summarizes the clinicopathologic features of colorectal epithelial neoplasms treated by ESD. Even small tumors 6 mm in size were resected by ESD because of the existence of scarring that rendered them non-lifting and difficult to resect with conventional EMR. Twenty-six tumors (13%) were histologically revealed to be invasive carcinoma. Seven tumors (3.5%), 6 with SM2 or deeper invasion including 2 tumors with vessel infiltration and 1 tumor with SM1 invasion Table 1. Clinicopathologic Features of Colorectal Epithelial Neoplasms Mean size (range) 29.9 mm (6 100) Location C/A/T/D/S/R 20/39/36/14/39/52 Macroscopic type 0-I 21 LST-G-H 45 LST-G-M 46 LST-NG-F 52 LST-NG-PD 18 SCAR 18 Histologic depth a Low-grade adenoma 44 High-grade adenoma 58 Noninvasive carcinoma 72 SM1 20 SM2 or deeper 6 Vessel infiltration Presence 3 Absence 197 C, cecum; A, ascending colon; T, transverse colon; D, descending colon; S, sigmoid colon; R, rectum; 0-I, protruding tumor. a The cut-off limit between SM1 and SM2 is 1000 m.

4 June 2007 COLORECTAL ENDOSCOPIC SUBMUCOSAL DISSECTION 681 Table 2. Features of Colorectal Epithelial Neoplasms According to Macroscopic Type 0-I LST-G-H LST-G-M LST-NG-F LST-NG-PD SCAR Mean size 28.8 mm 32.3 mm 42.2 mm 22.3 mm 21.1 mm 23.9 mm Histologic type & depth, % (no.) Low-grade adenoma 19.0 (4/21) 24.4 (11/45) 8.7 (4/46) 30.8 (16/52) 22.2 (4/18) 27.8 (5/18) High-grade adenoma 14.3 (3/21) 28.9 (13/45) 39.1 (18/46) 32.7 (17/52) 5.6 (1/18) 33.3 (6/18) Noninvasive carcinoma 33.3 (7/21) 46.7 (21/45) 37.0 (17/46) 30.8 (16/52) 27.8 (5/18) 33.3 (6/18) SM (5/21) 0 (0/45) 10.9 (5/46) 1.9 (1/52) 44.4 (8/18) 5.6 (1/18) SM2 or deeper 9.5 (2/21) 0 (0/45) 4.3 (2/46) 3.8 (2/52) 0 (0/18) 0 (0/18) Location, % (no.) Proximal colon 23.8 (5/21) 64.4 (29/45) 30.4 (14/46) 50.0 (26/52) 66.7 (12/18) 50.0 (9/18) Distal colon 23.8 (5/21) 11.1 (5/45) 21.7 (10/46) 38.5 (20/52) 33.3 (6/18) 38.9 (7/18) Rectum 52.4 (11/21) 24.4 (11/45) 47.8 (22/46) 11.5 (6/52) 0 (0/18) 11.1 (2/18) NOTE. The cut-off limit between SM1 and SM2 is 1000 m. 0-I, protruding tumor. plus vessel infiltration, were considered to be at substantial risk for nodal metastasis. Four of the 7 tumors were also treated by colorectal resection with lymphadenectomy. The other 3 tumors with possible nodal metastases were followed without additional treatment because the patients refused further treatment. From the analysis according to macroscopic type (Table 2), it was revealed that all LST-G-Hs and LST-NG-PDs were correctly diagnosed in terms of the depth of invasion, eg, slight submucosal invasion. Although the LST-G-Hs were all intramucosal tumors, nearly half of LST-NG-PDs (44%) were SM1 tumors, and one had vessel infiltration. Furthermore, type 0-I tumor and LST-G-M existed predominantly in the rectum, and the other tumor types existed predominantly in the proximal colon. There was no case of LST-NG-PD in the rectum. Overall rates of en bloc resection, en bloc plus R0 resection, R1 (lateral) resection, R1 (basal) resection, Rx (lateral) resection, and Rx (basal) resection were 91.5% (183/200), 70.5% (141/200), 18% (36/200), 0.5% (1/200), 11.5% (23/200), and 0% (0/200), respectively. Additional treatment was determined by the tumor depth and/or vessel infiltration and occurred only for the above 4 tumors that were also treated with colorectal resection with lymphadenectomy, including one tumor with SM2 or deeper invasion that was resected by R1 (basal) resection and R1 (lateral) resection. The technical outcomes of colorectal ESD according to macroscopic type are shown in Table 3. There was no significant difference in the rates of en bloc resection and en bloc plus R0 resection or complication rates according to macroscopic type. Perforation during ESD occurred in 11 lesions (5.5%), which were managed with conservative medical treatment after endoscopic closure of the perforation. One lesion (0.5%) resulted in a delayed perforation 2 days after the ESD, which was rescued by laparotomy. This patient had several comorbidities including hypertension, diabetes mellitus, post-sigmoidectomy and post left nephrectomy status, and chronic renal failure with hemodialysis. Minor bleeding was encountered in all the resections, but hemostasis was achieved during the procedures. No patient had massive hemorrhage requiring blood transfusion. Two resections (1%) in the rectum required emergency colonoscopies as a result of hematochezia after ESD to apply endoclips for causal vessels on the post-procedure ulcers. This was done on the same day as the ESD and 10 days after ESD, respectively. In 111 tumors (54 adenomas, 42 intramucosal carcinomas, and 15 SM1 carcinomas), we successfully obtained information about local recurrence through colonoscopy, which was performed more than 12 months after ESD. The median duration between ESD and the final colonoscopy was 18 months (range, months). The rates of en bloc resection, en bloc plus R0 resection, Rx (lateral) resection, and R1 (lateral) resection in the Table 3. Technical Outcomes of Colorectal ESD According to Macroscopic Type 0-I, % (no.) LST-G-H, % (no.) LST-G-M a, % (no.) LST-NG-F, % (no.) LST-NG-PD, % (no.) SCAR, % (no.) En bloc resection 95.2 (20/21) 88.9 (40/45) 93.5 (43/46) 94.2 (49/52) 88.9 (16/18) 83.3 (15/18) Resectability R (16/21) 66.7 (30/45) 65.2 (30/46) 76.9 (40/52) 72.2 (13/18) 66.7 (12/18) R1 (lateral) 9.5 (2/21) 20.0 (9/45) 21.7 (10/46) 17.3 (9/52) 16.7 (3/18) 16.7 (3/18) R1 (basal) 0 (0/21) 0 (0/45) 2.2 (1/46) a 0 (0/52) 0 (0/18) 0 (0/18) Rx (lateral) 14.3 (3/21) 13.3 (6/45) 13.0 (6/46) 5.8 (3/52) 11.1 (2/18) 16.7 (3/18) Rx (basal) 0 (0/21) 0 (0/45) 0 (0/46) 0 (0/52) 0 (0/18) 0 (0/18) Complication Perforation 4.8 (1/21) 8.9 (4/45) 8.7 (4/46) 3.8 (1/52) b 0 (0/18) 11.1 (2/18) Bleeding 0 (0/21) 0 (0/45) 4.3 (2/46) 0 (0/52) 0 (0/18) 0 (0/18) 0-I, protruding tumor. a This lesion is also resected by lateral positive margin. b This is a case of delayed perforation.

5 682 FUJISHIRO ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 5, No. 6 follow-up group were 93.7% (104/111), 69.4% (77/111), 10.8% (12/111), and 19.8% (22/111), respectively. Local recurrence was detected in 2 cases (1.8%) with multi-piece resections. The first case was a type LST-G-M noninvasive carcinoma, 5 cm in size, in high-grade adenoma located in the lower rectum, which was resected in 7 pieces with an R1 (lateral) resection. At the follow-up colonoscopy 2 months after ESD, the tumor recurred as a small adenoma on the ESD scar. This was treated by argon plasma coagulation, which resulted in no further recurrence. The other case was a type LST-G-M, 3 cm in size, SM1 recurrent carcinoma after repeated EMRs located in the transverse colon, which was resected in 3 pieces with an Rx (lateral) resection. At the follow-up colonoscopy 21 months after ESD, the tumor recurred as an SM1 carcinoma on the ESD scar, which was resected by partial colectomy. Survival data were obtained from 77 patients with malignant colorectal tumors (intramucosal carcinoma, 53; SM1 carcinoma, 18; SM2 or deeper carcinomas, 6) including 7 cases of possible nodal metastasis with or without additional colorectal resection. During a median follow-up period of 24 months (range, 6 74 months), one patient died from another coexisting malignant disease 23 months after ESD. All the other patients survived without evidence of distant recurrence after medical checkups performed by the doctors in charge. The 3 patients with tumors at high risk for nodal metastases and who refused any further surgical treatment were also recurrence-free at the follow-up periods of 10, 11, and 18 months after ESD. Discussion Several clinical studies regarding the efficacy and safety of ESD in the stomach have been reported. The technique shows high tumor eradication rates but substantial risks during the procedure. 6 We previously reported the possibility of using ESD in the esophagus 21 and the rectum, 7 which showed results similar to those of the stomach case series. 6 These findings suggest a high likelihood of successful application in the whole colorectum, but the key issue might be how to innovate in the procedures that were suitable for the colorectum. The primary consideration is reinforcing bowel cleansing. When the bowel is not well-prepared, the residual feces in the whole colorectum can be washed away by an endoscope equipped with a water-jet system before beginning ESD. The next consideration is technical refinement. The thinner gut wall can be easily perforated, so it is important to make a sufficient submucosal fluid cushion and to use cutting devices in an open field of view. From our previous study, we found a mixture of high molecular-weight hyaluronic acid, glycerin, and sugar useful as a submucosal injection solution. 22 Furthermore, the insulation-tipped electrosurgical knife commonly used in the stomach ESD 23 is discarded in the colorectum because a blade, not an edge, is used for cutting with the insulationtipped knife. We invented a new knife, the Flexknife. Cutting with a blade is sometimes used in a blind view field and leads to the cutting of an unexpectedly larger incision, which might result in perforation. The device with an edge as the cutting part can be applied in an open view field when used with a transparent attachment on the tip of an endoscope. 10,11 These considerations might result in immediate notification of perforation during ESD and successful management of the perforation, although the number of perforations was not less than that observed in the stomach case series. The cause of 1 delayed perforation with surgical rescue in our study is still unknown. Excessive thermal injury during ESD might have been given in this case, although the technique went smoothly and was similar to the other cases. However, we have to keep in mind that delayed perforation might occur after ESD, especially in patients who have several concurrent diseases. In our case series of colorectal ESD, a high en bloc resection rate and low recurrence rate were obtained. However, the rate of complications was still considerably high, even when the ESD was performed by surgeons who were very experienced with stomach ESD, and refinement of the technique was achieved. Considering that the technique was highly endoscopist-dependent, the tumors we indicated for ESD might not be suitable targets for ESD at other institutions. Subanalysis according to macroscopic type revealed that the targeted lesions in this study could be divided into 4 different groups: low and high malignant potential groups with complete preoperative prediction of node-negative tumors in terms of tumor depth; a group with incomplete preoperative prediction of node-negative tumors in terms of tumor depth; and a group that presented technical difficulties as a result of submucosal fibrosis. LST-G-H might be classified into low malignant potential groups with complete preoperative prediction of node-negative tumors from our case series, and a similar efficacy with a lower complication rate for LST-G-H was reported by using piecemeal EMR in a previous study. 24 Hence, another endoscopic option, such as piecemeal EMR, might be considered for LST-G-H. On the other hand, the high malignant potential for LST-NG-PD might be a sufficient reason to treat it in an en bloc fashion so as not to leave tumor cells in the submucosa, although we could predict that the tumor depth was limited to slight submucosal invasion in all the LST-NG-PDs we treated. The existence of an LST-NG-PD with vessel infiltration might require precise histopathologic analysis. For these reasons, ESD might become the first-line treatment for LST-NG-PDs. It might be permissible to treat a group with incomplete preoperative prediction of node-negative tumors in terms of tumor depth (type 0-I, LST-G-M, and LST-NG-F) endoscopically, because the majority of these can be curatively endoscopically resected. But a single-piece resection with attachment of a sufficient submucosal layer at the deepest level of tumor invasion is essential to precisely diagnose tumor depth in this group. Concerning this, a type 0-I tumor and LST-NG-F might be resected in an en bloc fashion because it is impossible to predict the deepest level of tumor invasion in both tumors, and submucosal invasion at multiple sites is reported in the latter tumor. On the contrary, it is reported that the deepest level of tumor invasion is predictable in LST-G-Ms because submucosal invasion is observed under a large nodular change. 25 Comparative studies are needed to determine an appropriate endoscopic treatment for these lesions. For example, ESD might achieve the en bloc resection of an entire lesion versus a piecemeal EMR, which might achieve single-piece resection at the deepest part of tumor invasion with less certainty. Finally, the treatment strategy for non-lifting lesions should be considered differently from the above, because these lesions contain various malignant potential, and it is technically impossible or risky to perform a conventional EMR technique. Before the advent of ESD, non-lifting intraepithelial neoplasms diagnosed endoscopically with a suspicion of malignancy were referred to a surgeon, and those unlikely to be malignant

6 June 2007 COLORECTAL ENDOSCOPIC SUBMUCOSAL DISSECTION 683 were treated by endoscopic ablation. ESD might be used as an alternative for the former lesions as a promising organpreserving treatment, although because of the relatively high perforation rate, the technique should be refined before it becomes standard procedure. In summary, this study showed that ESD is a promising technique for the resection of colorectal epithelial neoplasms. However, further evaluation and assessments of case series are necessary before ESD can be widely accepted as a standard endoscopic treatment in the colorectum. The benefit and underlying risk might be different, depending on the tumor s characteristics and the endoscopist s skill. Presently it might be better to resect certain subgroups of large, flat colorectal epithelial neoplasms or non-lifting epithelial neoplasms with submucosal fibrosis by using other options, such as piecemeal EMR or colorectal resection. Further studies are needed to confirm our results for colorectal ESD. References 1. Rembacken BJ, Fujii T, Cairns A, et al. Flat and depressed colonic neoplasms: a prospective study of 1000 colonoscopies in the UK. Lancet 2000;355: Hurlstone DP, Brown S, Cross SS. The role of flat and depressed colorectal lesions in colorectal carcinogenesis: new insights from clinicopathological findings in high-magnification chromoscopic colonoscopy. Histopathology 2003;43: Kudo S, Tamegai Y, Yamano H, et al. Endoscopic mucosal resection of the colon: the Japanese technique. Gastrointest Endosc Clin N Am 2001;11: Iishi H, Tatsuta M, Iseki K, et al. Endoscopic piecemeal resection with submucosal saline injection of large sessile colorectal polyps. Gastrointest Endosc 2000;51: Tamura S, Nakajo K, Yokoyama Y, et al. Evaluation of endoscopic mucosal resection for laterally spreading rectal tumors. Endoscopy 2004;36: Fujishiro M. Endoscopic submucosal dissection for stomach neoplasms. World J Gastroenterol 2006;12: Fujishiro M, Yahagi N, Nakamura M, et al. Endoscopic submucosal dissection for rectal epithelial neoplasia. Endoscopy 2006; 38: Participants in the Paris workshop. The Paris endoscopic classification of superficial neoplastic lesions: esophagus, stomach, and colon November 30 to December 1, Gastrointest Endosc 2003;58(Suppl):S3 S Kudo S, Shimoda R, Kashida H, et al. Laterally spreading tumor of colon: definition and history (in Japanese with English abstract). Stomach and Intestine 2005;40: Kodashima S, Fujishiro M, Yahagi N, et al. Endoscopic submucosal dissection using flexknife. J Clin Gastroenterol 2006;40: Kodashima S, Fujishiro M, Yahagi N, et al. Endoscopic submucosal dissection for gastric neoplasia: experience with the flex-knife. Acta Gastroentero Belg 2006;69: Fujishiro M, Yahagi N, Kashimura K, et al. Comparison of various submucosal injection solutions for maintaining mucosal elevation during endoscopic mucosal resection. Endoscopy 2004;36: Fujishiro M, Yahagi N, Kashimura K, et al. Different mixtures of sodium hyaluronate and their ability to create submucosal fluid cushions for endoscopic mucosal resection. Endoscopy 2004; 36: Yahagi N, Fujishiro M, Kakushima N, et al. Endoscopic submucosal dissection for early gastric cancer using the tip of an electro surgical snare (thin type). Digestive Endoscopy 2004; 16: Yahagi N, Fujishiro M, Imagawa A, et al. Endoscopic submucosal dissection for the reliable en bloc resection of colorectal mucosal tumors. Digestive Endoscopy 2004;16(Suppl):S89 S Oyama T, Tomori A, Hotta K, et al. Endoscopic submucosal dissection of early esophageal cancer. Clin Gastroenterol Hepatol 2005;3(Suppl):S67 S Fujishiro M, Yahagi N, Kakushima N, et al. Management of bleeding concerning endoscopic submucosal dissection with the flex knife for stomach neoplasm. Digestive Endoscopy 2006; 18(Suppl):S119 S Schlemper RJ, Riddell RH, Kato Y, et al. The Vienna classification of gastrointestinal epithelial neoplasia. Gut 2000;47: Dixon MF. Gastrointestinal epithelial neoplasia: Vienna revisited. Gut 2002;51: 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: Fujishiro M, Yahagi N, Kakushima N, et al. Endoscopic submucosal dissection of esophageal squamous cell neoplasms. Clin Gastroenterol Hepatol 2006;4: Fujishiro M, Yahagi N, Nakamura M, et al. Successful treatment outcomes of a novel endoscopic treatment for GI tumors: endoscopic submucosal dissection with a mixture of high-molecularweight hyaluronic acid, glycerin, and sugar. Gastrointest Endosc 2006;63: Ono H, Kondo H, Gotoda T, et al. Endoscopic mucosal resection for treatment of early gastric cancer. Gut 2001;48: Tanaka S, Haruma K, Oka S, et al. Clinicopathologic features and endoscopic treatment of superficially spreading colorectal neoplasms larger than 20 mm. Gastrointest Endosc 2001;54: Uraoka T, Saito Y, Matsuda T, et al. Endoscopic indications for endoscopic mucosal resection of laterally spreading tumors in the colorectum. Gut 2006;55: Address requests for reprints to: Mitsuhiro Fujishiro, MD, PhD, Department of Gastroenterology, Graduate School of Medicine, University of Tokyo, 7-3-1, Hongo, Bunkyo-ku, Tokyo, Japan. mtfujishkkr@umin.ac.jp; fax: