EMR of large, sessile, sporadic nonampullary duodenal adenomas: technical aspects and long-term outcome (with videos)

Transcription

1 ORIGINAL ARTICLE: Clinical Enoscopy EMR of large, sessile, sporaic nonampullary uoenal aenomas: technical aspects an long-term outcome (with vieos) Sina Alexaner, MBBS, FRACP, Michael J. Bourke, MBBS, FRACP, Stephen J. Williams, MBBS, FRACP, Aam Bailey, MB, ChB, FRACP, Jonar Co, MD Syney, Australia Backgroun: EMR is a viable alternative to surgery for removal of large mucosal neoplastic lesions of the entire GI tract. Few stuies have, however, been publishe on the safety, efficacy, an technical aspects of EMR in the uoenum. Objective: Our purpose was to evaluate the efficacy an safety of EMR of large (O15 mm) uoenal aenomas. Design: Retrospective evaluation of a efine patient cohort. Setting: Tertiary acaemic referral center. Patients: Patients with large (O15 mm) sporaic nonampullary uoenal aenomas manage by a stanarize technique who were referre by other specialist enoscopists for enoscopic treatment. Methos: Five-year ata from patients unergoing EMR for large uoenal aenomas were reviewe retrospectively. Immeiate an elaye complications were recore. Results: Twenty-one lesions were remove by EMR in 23 patients (mean age 62.2 years, 13 female, 10 male). The mean size of lesions resecte was 27.6 mm (meian 20 mm, range mm). Post-EMR histologic examination reveale mucosal aenocarcinoma in 1, low-grae tubulovillous aenoma (TVA) in 16, high- or focal high-grae TVA in 3 patients, an 1 with both high-grae TVA an carcinoi. EMR was performe successfully in 18 patients uring a single session. Two patients require 2 sessions an 1 require 3 sessions for complete resection. The meian follow-up was 13 months (range 4-44 months). During follow-up, 5 patients ha minor resiual aenomas that were treate successfully with snare resection an/or argon plasma coagulation. One patient ha EMR site bleeing. There were no perforations. Limitation: Retrospective stuy. Conclusion: EMR for large sporaic nonampullary uoenal aenomas is a safe an effective technique. (Gastrointest Enosc 2009;69:66-73.) Sporaic nonampullary uoenal aenomas are extremely uncommon. 1 The optimal treatment technique for large aenomas has not yet been efine. The majority of the publishe literature escribes the experience with surgical techniques. Enoscopic resection by EMR of a nonampullary uoenal lesion was first escribe in 1992, but since then there have been few reports of enoscopic resection in the uoenum. 2-4 The technique has not been stanarize, an ata on long-term outcomes are lacking. The aim of this stuy was to evaluate the short- an long-term outcomes of EMR of large, sessile, sporaic nonampullary aenomas in the uoenum. Abbreviations: APC, argon plasma coagulation; TVA, tubulovillous aenoma. DISCLOSURE: The authors report that there are no isclosures relevant to this publication. Copyright ª 2009 by the American Society for Gastrointestinal Enoscopy /$36.00 oi: /j.gie PATIENTS AND METHODS The stuy was approve by the human ethics an research committee. The enoscopy atabase of a tertiary referral, university-affiliate hospital was reviewe to ientify all patients with large nonampullary sessile uoenal lesions encountere over the preceing 5 years to December GASTROINTESTINAL ENDOSCOPY Volume 69, No. 1 :

2 Alexaner et al EMR of uoenal aenomas Patient files an enoscopy reports were reviewe, an where necessary structure telephone interviews were performe. Data collecte inclue lesion size an morphologic grae, histologic iagnosis, number of treatments, complications, an results of follow-up enoscopies. Only sessile lesions greater than 15 mm were inclue. Patients with polyposis synromes were exclue. Proceures were performe with the patient uner seation with a combination of miazolam, fentanyl, an propofol. Intravenous hyoscine 10 to 20 mg (Buscopan, Boehringer Ingelheim, Syney, Australia) was given at the time of EMR when require to limit uoenal motility. The iameter of the polyp was estimate at the time of enoscopy by reference to an open polypectomy snare place ajacent to the polyp. EUS was not routinely performe on these lesions. Instea, resectability was assesse at the time of the therapeutic proceure as etaile below. Technique All uoenal EMR proceures were performe by 2 experience interventional enoscopists (M. J. B. an S. J. W.) using a stanarize approach similar to that formally aopte for colon EMR in our center. 5,6 The relationship between the lesion an the papilla was routinely assesse, incluing use of a sie-viewing instrument if necessary. For lesions preominantly occupying the anteromeial wall of the uoenum, a uoenoscope was use, an for lesions on the lateral or posterior wall a peiatric variable-stiffness colonoscope was use (Olympus Optical, Tokyo, Japan). EMR was performe by the well-recognize general principles of tangential submucosal injection of a large volume (O20 ml) normal saline solution in combination with ilute epinephrine (1 in 10,000) an inigo carmine (0.04%, Mayne Pharma, Melbourne, Australia). 7,8 Freehan en bloc resection was then attempte for lesions of 15 to 20 mm (Figs. 1 an 2). Sequential piecemeal resection was performe for lesions greater than 20 mm with use of a combination of stiff-type snares, epening on lesion size an morphologic characteristics (20 mm SnareMaster, Olympus, Tokyo, Japan, or 15 mm 30 mm AcuSnare [mini oval], 25 mm 55 mm AcuSnare [stanar oval], Cook Meical, Brisbane, Australia). For carpet-like lesions, the serrate snare (SnareMaster) was preferre to enhance tissue capture. The mini oval snare was use to remove any resiual at the margin. For smaller or noular lesions, the mini oval or stanar oval snare was preferre. The technique inclue the following: Minimizing the number of separate injections to ecrease the likelihoo of transmucosal flui escape, which ha been our anecotal experience when multiple separate injection sites ha been use (in excess of 5). In cases of piecemeal resection, commencing at one lateral margin an generally proximally, incluing a small 1- to 3-mm margin of normal mucosal tissue (Figs. 3 an 4). The polyp capture was performe by pushing own firmly with an open snare on the target tissue while Capsule Summary What is alreay known on this topic Duoenal aenomas are usually solitary, but they have a rate of malignant transformation ranging from 35% to 85%. Enoscopic resection of large, sessile nonampullary uoenal aenomas has been consiere a high-risk proceure because of the thin uoenal wall. What this stuy as to our knowlege In a retrospective analysis of 21 patients who unerwent EMR for sessile, sporaic nonampullary uoenal aenomas greater than 15 mm, 18 lesions were remove in a single session, with 3 requiring multiple sessions for complete resection. At a meian follow-up of 13 months, 5 lesions resecte piecemeal ha remnant aenoma, but no recurrences were etecte among patients treate with en bloc resection. aspirating air as the snare was close. Reinsufflation was then performe, an appropriate positioning of the snare was visually confirme. In aition, the mobility of the mucosa relative to the muscularis propria was assesse by moving the snare back an forth several times. This was one for the ual purpose of ensuring that eeper layers ha not been ensnare an that the lesion was freely mobile relative to the muscularis propria, making eep invasion unlikely (Vieo 1, available online at www. giejournal.org). Recognition of the critical role of the first snare resection to arrive in the correct plane as evience by a significant wie mucosal efect with a blue base. Once in the correct plane, tissue was remove sequentially from that point, always working within that initial tissue plane, analogous to surgical issection. Attempt to resect the lesion in as few pieces as safely possible. For lesions that were greater than 40 mm or that occupie more than one wall of the uoenum, injection an resection of each wall or half segment was performe in turn to allow free access to the lesion an to enhance visualization (too large an initial injection may obscure the enoscopic view) an to minimize time for subsience of the cushion from flui iffusion. Two ifferent electrosurgical generators were use uring the stuy. Before 2005, a Soring generator (Arco- 3000, Quickborn, Germany), with a blene current of 60 watts cut an 30 watts coagulation, was use in 5 cases. After 2005, an ERBE electrosurgical unit (VIO 300, ERBE, Tubingen, Germany) set to Enocut Q, Effect 3, elivering a cut uration of 2 millisecons an a cut interval of 1200 millisecons was use. Volume 69, No. 1 : 2009 GASTROINTESTINAL ENDOSCOPY 67

3 EMR of uoenal aenomas Alexaner et al Figure 1. A, A 15-mm flat lesion on the lateral wall, with central epression. B, Complete en bloc snare resection with clean margins. Figure 2. A, A 20-mm lesion on the posterior wall in a patient on obligatory antiplatelet therapy (rug-eluting coronary stent). B, After en bloc resection. C, Defect close with 2 clips. Postproceural care an surveillance Mucosal resection proceures were performe on outpatients between 8 an 10 AM. After an uncomplicate stanar postenoscopy recovery, the patient remaine in secon-stage recovery for 4 to 6 hours on a clear liqui iet before final review by the proceuralist before ischarge. If the patients were well at that point, they were allowe home on a clear liqui iet until the following morning, after which a normal iet was to be resume. Written information was provie for the patient an the family or carers on potential problems an the signs along with a contact number. All patients receive twiceaily proton pump inhibitory therapy for 2 weeks after resection. Every attempt was mae to completely resect the lesion in a single session. Patients who ha unergone 68 GASTROINTESTINAL ENDOSCOPY Volume 69, No. 1 :

4 Alexaner et al EMR of uoenal aenomas Figure 3. A, A 60-mm carpet-like lesion occupying two thirs of the uoenum. B, After piecemeal resection. Figure 4. A, A 50-mm carpet-like lesion occupying one half of the uoenum. B, Defect after piecemeal resection. Minor venous oozing. a technically successful an complete resection with enoscopically clear margins were scheule for follow-up examinations in our unit at 3 an 12 months an then were avise to atten annually with the referring enoscopist. In cases where small amounts of resiual aenoma were etecte, these were resecte with the mini oval snare without submucosal injection. Argon plasma coagulation (APC) was use if necessary but only to treat minute resiuals that coul not be remove with the snare. In cases where only a scar was etecte, biopsies were performe for histologic confirmation. Histologic assessment All tissue was retrieve for histologic analysis, an in the case of multiple fragments a isposable retrieval net was use (Roth retrieval basket, Enome, Brisbane, Australia). All lesions resecte en bloc, an the ominant parts of larger polyps resecte piecemeal, were flattene an fixe with thin neeles onto cork before pathologic fixation. All specimens were reviewe by specialist GI pathologists. Lesions with high-grae ysplasia or invasive carcinoma were reviewe in conjunction with the investigators. RESULTS We ientifie a total of 23 patients with large, sessile, sporaic nonampullary uoenal aenomas. Of these, 21 unerwent EMR. All patients ha been referre by 19 specialist enoscopists from outsie our service. Baseline patient an polyp characteristics are shown in Table 1. The mean lesion size was 27.6 mm (meian 20 mm, range mm). Two polyps were locate at the junction of the first an secon parts, an 19 were in the secon part. EMR was performe on 21 lesions; 8 en bloc resection (mean size 15.6 mm) an 13 piecemeal (mean size 34.6 mm). All patients complete the enoscopic surveillance protocol. Two patients with 60- an 40-mm lesions ha enoscopic features suggestive for invasion, incluing nonlifting sign an lack of mobility when manipulate with the snare; thus EMR was not attempte. In 1 patient EUS was suspicious for submucosal invasion, an the patient unerwent surgery with eep submucosal invasion confirme. The secon patient (with a history of atrial fibrillation) ha a major cerebral ischemic event before EUS coul be performe, an EMR was not attempte. Volume 69, No. 1 : 2009 GASTROINTESTINAL ENDOSCOPY 69

5 EMR of uoenal aenomas Alexaner et al TABLE 1. Baseline patient an polyp characteristics (n Z 23) Mean age (y SD) Female/male ratio 13:10 Symptoms Iron eficiency anemia 3 Abominal pain 2 Nausea an vomiting 2 No symptoms 16 Figure 6. Abrupt termination of the mucosal fols in the scar at the site of resection of lesion seen in Figure 4A. Figure 5. Scar site 3 months after resection of the lesion shown in Figure 3A. Note the prominent villi. Histologic assessment after resection reveale mucosal aenocarcinoma in 1, low-grae tubulovillous aenoma (TVA) in 16, high or focal high-grae TVA in 3 patients an 1 with both high-grae TVA an a small 5-mm submucosal carcinoi that ha been completely excise. EMR was performe successfully in 18 patients uring a single session. Two patients require 2 sessions, an 1 require 3 sessions for complete resection. All 3 ha extensive lesions occupying more than two thirs of the uoenal circumference, an one ha also ha multiple prior attempts at resection before referral with resultant submucosal fibrosis. Aitional sessions were performe at 4 weekly intervals. Clinical an enoscopic follow-up is available in 20 an pening in 1 patient. Sixteen patients have unergone scheule examinations at 3 an 12 months, an 4 patients have complete 3 months of follow-up. The meian follow-up is 13 months (range 4 to 44 months). In 15 patients, enoscopic examination at 3 months reveale regenerate mucosa, often isplaying prominent villi (Fig. 5) an absence of the valves of kerckring an other mucosal fols in the scar (Fig. 6) with luminal narrowing in cases of extensive resection. Biopsy specimens from the sites in all patients were normal. In 5 patients (23.8%), minimal remnant tissue was seen, an all ha evelope in lesions that were resecte piecemeal, ranging from 15 to 50 mm in size. Three were treate successfully with snare resection in combination with APC, an two were treate with APC alone an were then clear at a subsequent enoscopy at 3 months. Meian follow-up is now 10 months in this group. There were no perforations or clinically significant immeiate bleeing. In 2 patients with a continuing nee for antiplatelet therapy for coronary stents, prophylactic closure of the entire efect (Fig. 2C) or the site of a visible vessel with clips was performe (Fig. 7 an Vieo 2, available online at One case of early bleeing evelope within 48 hours of the proceure an was treate successfully with enoscopic clip therapy. No bloo transfusions were require. One patient was amitte overnight with possible serositis an was treate with analgesia an intravenous antibiotics. An abominal CT scan showe normal results, an the patient was ischarge the following ay. DISCUSSION This stuy escribes the experience of a tertiary referral center in the enoscopic treatment of large, sessile, sporaic nonampullary uoenal aenomas with a stanarize EMR technique. Such lesions are not common, an their enoscopic resection has traitionally been consiere high risk because of the thin uoenal wall. 9 Aenomas in the uoenum are usually solitary an similar to colon lesions. However, they have a high rate of malignant transformation ranging from 35% to 85% Establishe treatment options are essentially surgical an inclue extensive uoenal segmental resection, submucosal excision after uoenotomy, or uoenopancreatectomy However, as the results of the current stuy show, EMR in the uoenum can be performe 70 GASTROINTESTINAL ENDOSCOPY Volume 69, No. 1 :

6 Alexaner et al EMR of uoenal aenomas Figure 7. A, A TVA occupying 50% of the circumference on the anterolateral wall in the istal secon part uoenum. B, Visible vessel in the right lateral aspect of the efect after resection. C, Two clips use to close the efect. safely an successfully in a tertiary referral setting on ayonly patients. Overall, remnant tissue at the site of prior resection was foun in 23.8% of our cases an occurre only in patients whose lesions were treate with piecemeal resection (5/13). All were subsequently treate with further snare resection in combination with APC or APC alone, an at a meian follow-up of 10 months they ha no evience of remnant aenomas. Our finings in the uoenum are not issimilar to those of the colon in which en bloc EMR of large lesions appears to be more effective in preventing recurrence. 19,20 Currently, lesions larger than 20 mm cannot safely be remove en bloc, an resiual remnant tissue remains a potential concern. Duoenal lesions resecte piecemeal require careful surveillance. The recurrence rate observe in this stuy is within the range of publishe ata for EMR of lesions elsewhere in the GI tract as well as surgically treate patients with uoenal aenomas. 14,19,21-23 Technique is critical in avoiing aenoma recurrence or resiual. Inclusion of a small margin of normal tissue in the lateral aspect of the resection may minimize the rate of recurrence an avoi the problem of small resiuals of neoplastic tissue at the margin of the efect. Also, careful sequential resection from the ege of the initial mucosal efect, working always from the ege of this avancing efect in combination with precise snare placement minimizes the potential for mucosal briges within the efect. These are ifficult to treat. These aspects of technique are not amenable to stuy within a ranomize trial setting but seem logical. However, further long-term follow-up stuies specifically examining this in the uoenum, an in particular using new technologies such as narrowban imaging or high-magnification chromoenoscopy at the time of EMR to examine the resecte margin, are require. It is likely that the safety an completeness of enoscopic resection of large uoenal aenomas is heavily epenent on the experience an expertise of the enoscopist. Apel et al 24 have escribe their experience in 18 patients with uoenal aenoma treate with snare polypectomy an APC accumulate over a 13-year perio. During a meian follow-up perio of 71 months, no patient ha carcinoma. Our preference is to resect first rather than ablate. In contrast to mucosal resection where a clean blue submucosal plane can be clearly iscerne, APC has no clear therapeutic en point other than aequate fulguration of resiual neoplastic tissue, the assessment of which is obfuscate by the accumulation of a coagulum at the site of therapy. Volume 69, No. 1 : 2009 GASTROINTESTINAL ENDOSCOPY 71

7 EMR of uoenal aenomas Alexaner et al When resiual aenoma is treate in follow-up, submucosal injection is generally avoie because this can create a canyoning effect from submucosal fibrosis beneath the target tissue. Submucosal injection in these circumstances leas to flui tracking laterally an preferentially elevating the normal mucosa aroun the target lesion but not beneath it. In this setting, the abnormal tissue is subsequently ifficult to capture with the snare. The peiatric variable colonoscope was preferre for posterior an lateral wall lesions because the working channel is oriente at the 6 o clock position. This facilitates firm ownwar pressure on the lesion, an the resection technique is very similar to that of colon EMR. If a gastroscope were use, the working channel exits at the 9 o clock position, an it is ifficult to simultaneously apply firm pressure with the snare from this position in aition to maintaining aequate visualization of the lesion. Thus, this instrument is generally avoie for the treatment of these lesions. The use of a sie-viewing instrument for the anteromeial uoenal wall allows en face viewing of the entire lesion. Such complete imaging is not always possible with a forwar-viewing instrument. For extensive lesions, on occasion both instruments are neee, or it may be necessary to inject submucosally to elevate the lesion into the fiel of view. Nonlifting or surface ulceration suggests submucosal invasion, an these lesions require further assessment with EUS an surgical treatment in the majority of cases. General guielines on enoscopic treatment of sessile lesions in the GI tract suggest that lesions occupying more than a thir of the luminal circumference shoul be consiere for surgery, but the current stuy inicates that in certain cases lesions occupying more than 60% of the circumference can be effectively treate enoscopically provie the enoscopist has goo access an the lesion elevates well with submucosal injection. Complications of EMR inclue perforation, bleeing, an serositis. Immeiate bleeing can often be controlle with enoscopic clip application. Because the uoenal wall is thin, it is our preference to anchor the clips on the ajacent normal mucosa to close the efect rather than clipping irectly into it. An early bleeing rate of up to 33% has been reporte in the uoenum, an it is the most common complication. 3 One limitation of this stuy is the retrospective esign with the potential for both the unerreporting of complications an selection bias. Selection bias has been minimize by incluing all patients with large, sessile, sporaic nonampullary uoenal aenomas ientifie within the atabase over the stuy perio. Because the ischarge instructions after EMR inclue etaile information on potential complications, an because a point of contact an enoscopic follow-up at 3 an 12 months was performe at our institution, we believe that the likelihoo of unerreporting of major complications is small. In conclusion, this stuy shows that the majority of lesions in the uoenum can be resecte uring a single outpatient proceure in a tertiary referral center with expertise in EMR technique. Careful enoscopic followup is essential to treat recurrence or resiual that, although rare in complete en bloc resections, can occur in giant lesions resecte piecemeal. REFERENCES 1. Murray MA, Zimmerman MJ, Ee HC. Sporaic uoenal aenoma is associate with colorectal neoplasia. Gut 2004;53: Obata S, Suenaga M, Araki K, et al. Use of strip biopsy in a case of early uoenal cancer. Enoscopy 1992;24: Ahma NA, Kochman ML, Long WB, et al. Efficacy, safety, an clinical outcomes of enoscopic mucosal resection: a stuy of 101 cases. Gastrointest Enosc 2002;55: Oka S, Tanaka S, Nagata S, et al. Clinicopathologic features an enoscopic resection of early primary nonampullary uoenal carcinoma. J Clin Gastroenterol 2003;37: Bourke MJ, Kaffes A, Ding S, et al. Safety, efficacy an long term follow up of chromoenoscopic saline cushion technique for enoscopic mucosal resection of large sessile colonic polyps [abstract]. J Gastroenterol Hepatol 2001;16:A Bourke MJ, Williams SJ, Gillespie PE, et al. Chromosaline enoscopic mucosal resection of large (O 20 mms) sessile colonic polyps; longterm follow up. J Gastroenterol Hepatol 2003;18:B50, Karita M, Taa M, Okita K, et al. Enoscopic therapy for early colon cancer: the strip biopsy resection technique. Gastrointest Enosc 1991;37: Kuo S, Kashia H, Nakajima T, et al. Enoscopic iagnosis an treatment of early colorectal cancer. Worl J Surg 1997;21: Ponchon T. Enoscopic mucosal resection. J Clin Gastroenterol 2001;32: Galaniuk S, Hermann RE, Jagelman DG, et al. Villous tumors of the uoenum. Ann Surg 1988;207: Miller JH, Gisvol JJ, Weilan LH, et al. Upper gastrointestinal tract: villous tumors. AJR Am J Roentgenol 1980;134: Perzin KH, Brige MF. Aenomas of the small intestine: a clinicopathologic review of 51 cases an a stuy of their relationship to carcinoma. Cancer 1981;48: Schulten MF, Oyasu R, Beal JM. Villous aenoma of the uoenum: a case report an review of the literature. Am J Surg 1976;132: Farnell MB, Sakorafas GH, Sarr MG, et al. Villous tumors of the uoenum: reappraisal of local vs. extene resection. J Gastrointest Surg 2000;4: Chappuis CW, Divincenti FC, Cohn I Jr. Villous tumors of the uoenum. Ann Surg 1989;209: Krukowski ZH, Ewen SW, Davison AI, et al. Operative management of tubulovillous neoplasms of the uoenum an ampulla. Br J Surg 1988;75: Scott-Coombes DM, Williamson RC. Surgical treatment of primary uoenal carcinoma: a personal series. Br J Surg 1994;81: Pezet D, Rotman N, Slim K, et al. Villous tumors of the uoenum: a retrospective stuy of 47 cases by the French Associations for Surgical Research. J Am Coll Surg 1995;180: Kaltenbach T, Frielan S, Maheshwari A, et al. Short- an long-term outcomes of stanarize EMR of nonpolypoi (flat an epresse) colorectal lesions R1 cm (with vieo). Gastrointest Enosc 2007;65: Hurlstone DP, Cross SS, Brown S, et al. A prospective evaluation of high-magnification chromoscopic colonoscopy in preicting completeness of EMR. Gastrointest Enosc 2004;59: GASTROINTESTINAL ENDOSCOPY Volume 69, No. 1 :

8 Alexaner et al EMR of uoenal aenomas 21. Iishi H, Tatsuta M, Iseki JK, et al. Enoscopic piecemeal resection with submucosal saline injection of large sessile colorectal polyps. Gastrointest Enosc 2000;51: Kanamori T, Itoh M, Yokoyama Y, et al. Injection-incision assiste snare resection of large sessile colorectal polyps. Gastrointest Enosc 1996;43: Noa M, Koama T, Atsumi M, et al. Possibilities an limitations of enoscopic resection for early gastric cancer. Enoscopy 1997;29: Apel D, Jakobs R, Spiethogg A, et al. Follow-up after enoscopic snare resection of uoenal aenomas. Enoscopy 2005;37: Receive February 26, Accepte April 19, Current affiliations: Department of Gastroenterology an Hepatology, Westmea Hospital, Syney, Australia. Reprint requests: Michael Bourke, MBBS, FRACP, Westmea Hospital, Enoscopy Unit, Hawkesbury Roa, Westmea, Syney, New South Wales, 2145, Australia. If you want to chat with an author of this article, you may contact him at Volume 69, No. 1 : 2009 GASTROINTESTINAL ENDOSCOPY 73