|Home | About | Journals | Submit | Contact Us | Français|
Correspondence to: Marc Barthet, MD, Professor, Department of Gastroenterology, Hôpital Nord, Chemin des Bourrely, 13915 Marseille cedex 20, France. marc.barthet/at/ap-hm.fr
Telephone: +33-491-968737 Fax: +33-491-961311
AIM: To evaluate the feasibility and the outcome of endoscopic mucosal resection (EMR) for large colorectal tumors exceeding 4 cm (LCRT) undergoing piecemeal resection.
METHODS: From January 2005 to April 2008, 146 digestive tumors larger than 2 cm were removed with the EMR technique in our department. Of these, 34 tumors were larger than 4 cm and piecemeal resection was carried out on 26 colorectal tumors. The mean age of the patients was 71 years. The mean follow-up duration was 12 mo.
RESULTS: LCRTs were located in the rectum, left colon, transverse colon and right colon in 58%, 15%, 4% and 23% of cases, respectively. All were sessile tumors larger than 4 cm with a mean size of 4.9 cm (4-10 cm). According to the Paris classification, 34% of the tumors were type Is, 58% type IIa, 4% type IIb and 4% type IIc. Pathological examination showed tubulous adenoma in 31%, tubulo-villous adenoma in 27%, villous adenoma in 42%, high-grade dysplasia in 38%, in situ carcinoma in 19% of the cases and mucosal carcinoma (m2) in 8% of the cases. The two cases (7.7%) of procedural bleeding that occurred were managed endoscopically and one small perforation was treated with clips. During follow-up, recurrence of the tumor occurred in three patients (12%), three of whom received endoscopic treatment.
CONCLUSION: EMR for tumors larger than 4 cm is a safe and effective procedure that could compete with endoscopic submucosal dissection, despite providing incomplete histological assessment.
Endoscopic mucosal resection (EMR) allows complete and curative removal of the affected mucosa by excising through the middle or the deeper part of the submucosa following isotonic saline injection[1-3]. Thanks to a very simple technique, the endoscopic resection of superficial colorectal adenomas and tumors has been made possible in selected patients with little or no ganglion risk. The therapeutic management of polyps larger than 3 cm, however, often relies on surgery. The frequency of degeneration is indeed increased, and piecemeal resection is carried out in 76% of cases. Laparoscopic surgery has reduced the length of hospitalization, but a substantial morbidity rate of 20% and a mortality rate of 1% persist. Endoscopic submucosal dissection (ESD) is a technique currently undergoing evaluation. This technique offers the advantage of a monobloc resection, enabling the analysis of the lateral margins and leading to a reduction of the risk of recurrence. However, this technique’s learning curve is steep, and the morbidity associated with it is higher than that of mucosectomy.
This study aimed to evaluate the feasibility and the outcome of endoscopic piecemeal mucosal resection of sessile polyps of the colorectum exceeding 4 cm in size.
Between 2005 and 2007, our department carried out 146 mucosal resections larger than 2 cm throughout the gastrointestinal tract. Of these, 34 involved parietal tumors larger than 4 cm, and 26 tumors were colorectal. These large colorectal tumors exceeding 4 cm were treated by piecemeal resection in 25 of the 26 patients. Among these, 44% were female and the mean age was 71 years (46-89 years).
The lesions were first identified by visual inspection (Figure (Figure1A).1A). Computed virtual chromoendoscopy with Fujinon Intelligent Color Enhancement® was used for estimating surface extension. The size of the lesions was measured and compared using open biopsy forceps. Deep tumor extension was estimated using three parameters. Polyps were classified according to the Paris morphological classification. All lesions were analyzed according to Kudo’s pit pattern classification. Finally, endosonography was performed on all rectal lesions. Only lesions classified as T1N0 were treated endoscopically.
Certain lesions were excluded from the study and treated by surgery if they met one or more of the following criteria: (1) Type 0-III, Paris classification; (2) Type V, Kudo’s classification; (3) Endosonographic lesions > T1 or N+; and (4) Absence of lifting after submucosal injection (negative lifting sign).
A video colonoscope type Fuji 450® was used for total colonoscopy, carried out under general anesthesia. Patients received endocarditis prophylaxis according to the recommendations of the Société Française d’Endoscopie Digestive. Mucosal resection was performed using the conventional method. In order to create a detachment of the pathological mucosa, a solution containing physiological serum (0.9%) and epinephrine (dilution of 1:10 000) was injected into the submucosa by means of a 25 gauge needle inserted into the operating channel (Figure (Figure1B).1B). A diathermic snare was tightened around the elevated lesion, grasping also adjacent healthy mucosal tissue (Figure (Figure1C).1C). Two snare types were used: large snare of 6 cm × 3 cm (Jumbo Wilson Cook®, USA) and needle snare of 5.5 cm × 2.5 cm (Wilson Cook®, USA). Removal was performed with pure cutting current in the caecum and with endocut current in the remaining colon (Figure (Figure1D).1D). Considering that the injection could have masked the polyp’s limits, the healthy peripheral mucosa was previously marked with a bistoury in order to ensure complete polyp removal. In case of doubt as to the normality of the resection margins, a complementary treatment was performed by applying a coagulation current to the mucosectomy margins, either via the tip of the snare or via argon plasma, at a flow rate of 1 L/min and a power of 60 W (Erbe®, Tuebingen, Germany). One-time resection was attempted in all patients (Figure (Figure1E).1E). If the mucosectomy was considered incomplete due to the persistence of residual adenomatous tissue, a further mucosectomy was planned.
Each fragment of removed tissue was spread out and pinned on a 5 mm thick cork surface in order to avoid retraction of the polyp’s base and permit improved histological assessment of the deep margins. Deep resection limits were marked with China ink and the tissue pieces were then fixed in 4% formol saline for 3 d. All fragments were routinely processed for paraffin embedding. Slices (5 μm) were stained with hematein-eosin-safran. Modified Vienna classification was used for the histological assessment according to the severity of dysplasia. Lateral margins could not be evaluated due to the impossibility of displacing fragments in relation to each other. Histological assessment was carried out using the Japanese classification with deep submucosal invasion limited to 1000 μm.
Following mucosal resection, colonoscopic surveillance included a first early endoscopic control examination between 3 and 6 mo, and further examinations at 1 and 3 years (Figure (Figure2).2). Resection was considered complete if no residual adenomatous tissue was noted following completion of mucosectomy. Local recurrence was defined as the presence (on biopsy) of adenomatous tissue in areas of previously treated mucosa at endoscopic control between 3 and 6 mo after therapy. In the case of a recurrence of small-sized tumors, an additional mucosectomy was carried out. If a lesion could not be lifted due to fibrous tissue development in the submucosa, argon plasma coagulation (APC) following multiple biopsies was performed. When the biopsy revealed recurrent cancer, the patient underwent surgery.
The 26 sessile polyps were removed by piecemeal resection. The mean polyp size was 4.9 cm (4-10 cm). Lesion analysis according to the Paris classification revealed the presence of type 0-Is in 9 cases (34%), type 0-IIa in 15 cases (58%) and type 0-IIb in one lesion (4%), with some mixed forms. The only lesion classified as type 0-IIc (4%) showed low-grade dysplasia and two lesions classified as type 0-IIa were intramucosal carcinomas. There was no ulcerated lesion of type 0-III. The most frequently identified pit pattern was pit pattern type IV in 24 cases (92%) followed by type III. The predominant location was the rectum in 15 cases (58%), and all lesions were classified as T1N0 (Table (Table11).
A one-time resection was performed in 23 cases (88.5%). For two lesions located in the rectum (7.7%), three sessions were necessary, and for one lesion located at the right colonic angle (3.8%), two sessions were needed. The lesions’ mean size was 6.7 cm (4-10 cm). Coagulation of margins was performed as complementary treatment in 38.4% of cases. The mean duration of the intervention was 65 min (25-137 min).
All polyps were adenomatous. The most common histological type was villous adenoma in 42% (11 cases), followed by tubular adenoma in 31% (8 cases) and finally tubulovillous adenoma with a mixed villous and tubular architecture in 27% (7 cases). In total, 35% of lesions exhibited low-grade dysplasia (9 cases), 38% high-grade dysplasia (10 cases), 19% a carcinoma in situ (5 cases). Two polyps (8%) were reported as intramucosal carcinoma (m2), neither of which reached the submucosa. Contrary to the lateral margins (0%), deep margin analysis was possible on all samples. Resection was performed within a safe deep margin in 100% of cases.
In total, 24 patients underwent endoscopic surveillance, while one patient refused to undergo control endoscopy. The mean duration of follow-up was 12 mo (3-37 mo). Three recurrences were detected (12.5%) (Table (Table2).2). All three patients had initially received a complementary treatment using APC. The median delay until recurrence diagnosis was 3 mo (3-9 mo). Two patients underwent endoscopic treatment and the remaining patient received surgical therapy. The first patient received an additional mucosectomy, followed by APC due to the absence of lifting after the injection. For the second patient, a single session of tissue destruction via coagulation was the immediate resort. These two endoscopically-treated recurrences were of the same histological type as the initial polyp. Following treatment of recurrences, a further endoscopic control was scheduled for 3 mo later. The third patient, who was followed-up for Crohn’s disease which was surgically treated by ileorectal anastomosis, underwent rectal stump resection. After three successive endoscopic treatments, one using mucosectomy and the two others using APC, the low-grade dysplastic lesion detected at biopsy could not be eradicated. Histological examination of the surgically removed specimen revealed an infiltrating adenocarcinoma that was classified as pT2N0 with a colloid mucosal component of less than 50%. In total, endoscopic resection was effective in 96% of cases.
Intra-operative bleeding occurred in two patients but in neither patient was there any need for blood transfusion or a drop in hemoglobin levels exceeding 2 g/dL. Late postoperative bleeding, at day 6, was observed in two patients (7.7%) but no case of severe bleeding as defined in the standards of practice of the American Society for Gastrointestinal Endoscopy was noted. Endoscopic injection therapy with epinephrine (diluted 1:10 000) was performed along with the deployment of hemostatic clips (Resolution® Boston Scientific, USA).
Perioperative perforation was diagnosed in one case (4%) upon visualisation of the peritoneal fat at the bottom of the resection zone. The abdominal radiography without preparation did not reveal a pneumoperitoneum because of the perforation’s sub-peritoneal rectal location. The patient underwent conservative treatment with endoscopic closure of the perforation using clips, in addition to the administration of antibiotics and 48 h fasting.
Scar stenosis was observed in another patient who presented a 10 cm circumferential lesion within the upper rectum. The interval between the first colonoscopy and the diagnosis of the stenosis was 6 mo. The patient was asymptomatic and the endoscopic CRE balloon dilatation was able to traverse the stenosis.
Ten days after his examination, one patient experienced septic shock following Staphylococcus epidermis endocarditis. The patient had a bicuspid aortic valve, and underwent prosthetic valve replacement surgery. The microbiology did not support a bacterial translocation from the digestive tract.
This study aimed to demonstrate the efficacy of piecemeal mucosectomy in the management of large non-polypoid colorectal lesions. The success rate was 96% despite a mean lesion size of 4.9 cm. These results are comparable to those reported in the literature, which show success rates of endoscopic treatment for large polyps (those exceeding 2 cm) ranging from 83% to 100% (Table (Table3).3). Kaltenbach et al recently described a series of 125 mucosectomy-treated colorectal plane lesions whose mean size was 16.7 mm. Of the 62 patients followed up for a mean period of 4.5 ± 1.4 years, none had developed colorectal cancer or metastasis, resulting in a success rate of 100%.
The Japanese Society for Cancer of the Colon and Rectum’s current criteria for curative endoscopic resection are: a submucosal invasion of less than 1000 μm, moderate or well-differentiated lesion characteristics and the absence of vascular invasion[14,15]. Depending on the study reported, the presence of invasive lesions, which cross the muscular mucosa, varies from 0% to 44%, notably for sessile elevated tumors[1,16]. In our series, there was no submucosal invasion and only 2 cases (8%) with invasion of the mucosa (m2). This was accounted for by the parietal extension assessment, which was based on a series of complementary examinations. On one hand, the Paris endoscopic classification allows the exclusion of ulcerated lesions. On the other hand, magnified chromoendoscopy has been shown to permit a 97% diagnostic precision of deep submucosal invasion (69/71). However, another study found a specificity of only 50%. Echoendoscopy with high-frequency mini probes was not used, as muscular mucosa is only visualized in 50% to 65% of cases[19,20]. Mini probes of 7.5 MHz were employed, permitting the detection of T1 cancers without separating mucosal and sub-mucosal lesions. The evaluation of extension depth is an essential step prior to resection in order to select the patients for whom endoscopic treatment will be curative with minimal lymphatic risk.
The late bleeding rate reported by Doniec et al was only 2%. The higher late bleeding rate observed in our study (7.7%) can be explained by the 49 mm mean extension of the resection. Lesion size has been identified as an independent predictive factor of postpolypectomy delayed bleeding. However, the bleeding complications of the current study and those aforementioned were effectively treated with the deployment of hemoclips and the injection of an adrenaline serum, without resorting to surgery. Perforation rates reported after the resection of polyps exceeding 2 cm have been shown to vary from 0% to 1.2%[23,24]. If the perforations are of small size (< 1 cm) and are properly diagnosed during endoscopy, they may be treated with endoscopic clip closure in 70% of cases. Scar stenosis of the colon is a complication following endoscopic mucosectomy which has not yet been explained. Two factors appear to play a role: the initial lesion’s extended circumferential shape, and its location in the upper rectum (where the diameter is smallest).
Local recurrence was detected in 13.4% of cases, despite APC of the area of the defect which has proved effective in preventing recurrences following piecemeal mucosal resection. Previous studies have reported recurrence rates ranging from 0% to 39% for lesions larger than 2 cm[22,27]. In the largest series, that of Seitz et al, 288 patients with large (> 3 cm) sessile and pediculated polyps were treated using piecemeal resection, and the recurrence rate was 17% for a mean follow-up duration of 36 mo. The mean delay for diagnosis of recurrence was 5 mo. We followed the recommendations of Higaki concerning postpolypectomy surveillance. Most recurrences occur within the first 6 mo. Our own experience and published data encourage close endoscopic control between 3 and 6 mo, and at 1 year, in cases of piecemeal resection. Patients who have undergone resection of a large dysplastic or cancerous polyp require tighter endoscopic control. Several risk factors for recurrence have been identified; piecemeal resection, lesion size, granulous appearance of the lesion, and a lesion location at the bottom of the rectum attaining the pectineal line[23,32]. Among these factors, endoscopic piecemeal procedure appears to play the most important role. In a series by Ishihara et al, 78 esophageal squamous cell carcinomas of ≥ 2 cm were treated by endoscopy. The strongest predictor for recurrence was the number of resected fragments; 0/34 in patients with monobloc resection (0%), 4/24 in those with 2 and 4 fragments (15%) and 8/17 in those with more than 5 fragments (47%). In addition to the risk of local recurrence, piecemeal resection may prevent a good assessment of the lateral margins, with the consequence of incomplete interpretation and the risk of leaving small carcinomas around the main lesion. In a recent retrospective study by Kim et al, 44 patients who had initially received imperfect EMR for colorectal cancers were subsequently treated by either EMR or ESD. Gross incomplete resection and deep margin positivity were found to be risk factors of residual cancer. No residual cancer cells were found after supplementary surgery in all cases with positive lateral resection margins. The authors proposed the hypothesis that the application of an electrocoagulation current can destroy residual cancer cells at the resection margins. Hurlstone et al have shown that recurrence rate could be decreased from 8.7% to 0.5% with the analysis of the pit patterns of the resection margins.
If the recurrence is of small size, removal by hot forceps is a therapeutic option. If the recurrence exceeds 5 mm, a further mucosectomy session may be carried out so as to obtain a complete histological analysis of residual tumor tissue. In our series, such a thorough histological analysis was not possible in two out of three cases. The thermal lesion provoked by the coagulation current leads to a desmoplasmic reaction in the submucosa, preventing another mucosal lifting (non lifting sign). Destruction of tumors by APC is an alternative therapeutic option. In a series of 68 recurrent colorectal adenomas following EMRP, APC treatment was effective in 90%. We believe that in all cases without malignant recurrence, the best choice is to try to perform a new EMR, in order to assess the margins. However, the submucosal fibrotic scar due to thermal injury may prevent the lifting sign leading to a management by local destruction, mainly with APC treatment. Surgery remains essential in the presence of a carcinomatous recurrence whose extension depth cannot be properly assessed, and which therefore has a risk of metastatic spread. In our series, one patient followed-up for Crohn’s disease with ileorectal anastomosis underwent surgery. The operative specimen revealed an infiltrating cancer classified as pT2N0. A recent meta-analysis has shown that the positive predictive value of a plane low-grade dysplastic lesion, associated with chronic inflammatory bowel disease, correlated with the presence of colorectal cancer in 22% of cases.
The technique of large EMR for rectal tumor might be challenged by two new techniques: ESD and transanal endoscopic microsurgery (TEMS). ESD permits single-fragment resection with safe margins in 70% of cases, as has been shown in a series of 200 colorectal lesions ranging in size from 20 to 150 mm. There are only a few follow-up studies, but the risk of recurrence appears low. In a study by Fujishiro et al, 35 rectal polyps were treated using the ESD method. Monobloc resection with safe margins was possible in 62.9% of cases, with a local recurrence rate of 2.8% for a mean follow-up duration of 36 mo. Initially developed for the treatment of superficial gastric tumors, the application of this technique in the colon exposes the patient to a higher risk of complications. For instance, ESD is associated with a higher risk of perforation, ranging from 5% to 14%, due to the thinness of the colonic wall[40,25]. The intra-operative bleeding risk is also significantly higher for ESD as compared to EMR (22.6% for ESD vs 7.6% for EMR, P < 0.01). The “inject and cut” mucosectomy has the advantage of being a simple technique that may be performed by the endoscopy practitioner with routinely-used equipment. In contrast, ESD practitioners require extensive training, operating initially on animals because the endoscopist’s experience inversely correlates with the risk of perforation and monobloc resection rate[43,44]. The mean time required for the procedure is significantly higher and has been compared in a study by Oka et al. The mean time for the resection of gastric lesions smaller than 1 cm was 3.5 ± 1.3 min for EMR and 58.5 ± 28.7 min for ESD (P < 0.01). If the lesion is larger than 2 cm, the mean time required for the procedure is 17.2 ± 9.3 min for EMR and 123.8 ± 101.4 min for ESD (P < 0.01). ESD is at present a non-standardized technique. It has the advantage of allowing monobloc resection of gastrointestinal tumors, which permits the assessment of lateral margins. However, the technique is difficult to learn, its complication rates are high and the time required for the procedure is long. To date, this technique cannot be proposed as a standard technique for the management of colorectal adenomas.
For TEMS, major complications have been reported. In a series of 288 patients, 9% of patients experienced complications; severe digestive hemorrhages, perforations, and temporary anal incontinence. Non-surgical complications accounted for the low but relevant mortality rate of 0.3%. Duration of hospitalization was approximately 2 d. Hurlstone et al have evaluated EMR as an alternative to TEMS. In 62 patients, the success rate of endoscopic treatment was 98%, with a recurrence rate of 8%. Patients were discharged the same day as the procedure in 97% of cases. Another major drawback of TEMS is economic, since it necessitates a financial investment of 50 000 Euros for the purchase of the equipment.
In conclusion, endoscopic piecemeal resection is a safe and effective procedure. In spite of higher recurrence rates than with the ESD method, its technical simplicity, low complication rates, and lower costs are major advantages. Moreover, endoscopic management of recurrence has proven effective, with no risk of submucosal invasion in patients without chronic inflammatory colitis. A large lesion size (> 4 cm) is not a limiting factor for an endoscopic approach, provided that the risk of submucosal invasion has been carefully evaluated. However, further studies are needed to specify the time intervals required for endoscopic surveillance and to develop new techniques that would allow the histological assessment of lateral margins.
Large colorectal adenomatous sessile polyps are neoplastic premalignant lesions, which carry a high risk of transformation into invasive cancer. The management of these lesions is usually surgical by segmental colonic resection. The treatment of these lesions by piecemeal mucosectomy represents a mini-invasive procedure and the resection is effective for selected lesions.
Endoscopic mucosectomy is a technique approved for lesions less than 2 cm because it allows en bloc resection. The endoscopic resection of lesions more than 4 cm involves a systematic piecemeal resection, which could result in the absence of anatomopathologic analysis of the margins and may expose the patient to the risk of recurrence.
This is an original study taking place in a single centre that describes the effectiveness and the feasibility of this technique as less risky and simpler than that of endoscopic submucosal dissection (ESD), for the management of polyps of large size (giant polyps). This study confirms that the size of the polyps should not be a contraindication for endoscopic treatment. However, the risk of deep parietal infiltration must be evaluated in a precise way because this controls the effectiveness of the treatment as well as the risk of metastatic dissemination to lymph nodes which could preclude the endoscopic resection. It is associated with a low risk of complications and recurrence.
The results of this study show that mucosectomy for large colorectal polyps is effective whatever the size of the lesions if the estimation of lymph node infiltration is rigorously evaluated by the degree of infiltration into the layers of the wall of the colon. However, new studies should be realized in order to confirm these data and to determine the degree and regularity of follow-up monitoring.
Mucosectomy is an endoscopic technique of resection of a lesion that requires the separation of the submucosa using normal saline solution. ESD is a new method of resection, allowing the dissection of the lesion within the thickness of the submucosa or the interface between the submucosa and the muscularis propria. FICE: Technique of virtual chromoendoscopy by processing the image in a narrow spectral band.
Great paper, well written, appeals to our audience, favor acceptance.
Peer reviewers: Chris JJ Mulder, Professor, Department of Gastroenterology, VU University Medical Center, PO Box 7057, 1007 MB Amsterdam, The Netherlands; Walter E Longo, Professor, Department of Surgery, Yale University School of Medicine, 205 Cedar Street, New Haven, CT 06510, United States
S- Editor Tian L L- Editor Logan S E- Editor Zheng XM