Management of Ampullary Adenomas
Q: What Is the Best Treatment for Duodenal or Ampullary Adenoma, and What Is the Current Recommendation for Surveillance After Treatment?
A: Ampullary tumors historically presented late, with jaundice, bleeding, relapsing pancreatitis, or duodenal obstruction, and were treated with radical pancreaticoduodenectomy and, less commonly, transduodenal resection and dual sphincteroplasty. Endoscopy in this setting was limited diagnostically to tissue acquisition and therapeutically to biliary stent placement or sphincterotomy in high surgical risk patients with obstructive jaundice. Currently, the majority of adenomas are found incidentally during screening for reflux or dyspepsia or follow-up of C-loop adenomas in patients with Gardner’s syndrome.
Earlier diagnosis allows the potential for endoscopic as opposed to surgical resection, but neither endoscopic papillectomy nor local tumor ablation should be considered in isolation. If found incidentally, multiple biopsies should be performed to rule out malignancy or a tumor with unusual histology such as carcinoid (Figure 1). Symptomatic patients, in turn, invariably require an abdominal computed tomography (CT) scan and measurement of tumor markers to include CEA and CA 19-9. Assuming that CT and biopsy fail to demonstrate malignancy, there have been a number of attempts to define endoscopic nonresectability. The latter include extremely large (4- to 5-cm) lesions, an ulcerated papilla, lesions that extend laterally and encompass greater than 30% to 50% of the duodenal lumen, and a hard lesion that fails to elevate with a submucosal injection. Likewise, polyps that extend a significant length (0.5 to 1.0 cm) into the bile duct or pancreas as demonstrated by endoscopic retrograde cholangiopancreatography (ERCP) or magnetic resonance cholangiopancreatography (MRCP) are considered by many to be unresectable endoscopically (Figure 2). The role that endoscopic ultrasound (EUS) or intraductal ultrasound (IDUS) plays in patients with ampullary adenomas is controversial. Some endoscopists believe that EUS should be undertaken in all patients with ampullary adenomas to preclude the risk of resection in patients with invasive malignancy. I use EUS selectively and reserve it for patients in whom pancreas protocol CT fails to show malignancy but in whom local infiltration or the question of endoscopic resectability remains uncertain.
Technically, with the exception of asymptomatic and nonenlarging microadenomas of the papilla of Vater (which are virtually ubiquitous in Gardner’s patients), ampullary adenomas should be treated. As noted previously, historical treatment was usually done surgically, either with transduodenal papillectomy with concomitant sphincteroplasty or with a Whipple procedure.
Endoscopic treatment of ampullary adenomas was initially limited to direct thermal ablation, most commonly with a Nd-YAG laser, but now is more commonly performed with argon plasma coagulation (APC). Both treatment modalities have been associated with edematous obstruction of the pancreaticobiliary outlet with resultant jaundice, cholangitis, or obstructive pancreatitis. As a consequence, these modalities are currently used primarily to treat residual adenomatous tissue at the edge of a papillectomy site. Snare papillectomy, in turn, can be done in a piecemeal fashion for lesions less than 2 to 3 cm but can usually be done with a single resection using a blended current. I do an ERCP in all patients prior to resection, to assure that there is no tumor growth into the pancreaticobiliary (PB) tree (see Figure 2), as well as to define landmarks to facilitate postpapillectomy PB drainage.
Some endoscopists utilize submucosal papilla injection to buffer the duodenal wall prior to resection, as well as to define a non-“lifting sign” in the setting of infiltrating neoplasm. I do not. Such injections often blur tissue planes and require considerably more electrical current to facilitate transection of the papilla.
Following baseline ERCP and papillectomy, I undertake a biliary sphincterotomy, using a blended or pulsed current, and pancreatic sphincterotomy with pure cut current (Figure 3), placing stents in both ducts to protect against edematous obstruction of the duct orifices (Figure 4). APC treatment of the edges or small bleeding areas is sometimes required. Data from a multicenter, retrospective series by Catalano et al have noted a 17% incidence of pancreatitis in patients in whom no stent was inserted vs 3.3% of patients undergoing prophylactic pancreatic duct stenting. Moreover, subsequent stenosis of the PD sphincterotomy fell from 18.4% of patients to 1.1% in patients who had prophylactic stenting. Although early and late biliary complications were comparable between patients who did or did not have biliary stenting, early bouts of cholangitis and cholestasis have changed my practice patterns to favor prosthesis placement in most patients.
There are now numerous series that suggest that thermal ablation of the papilla is associated with a higher side-effect profile, less effective eradication of the ampullary adenoma, and need for significantly more interventions when compared with papillectomy. Moreover, there appears to be a 3- to 4-fold increased risk for ultimate development of malignancy when compared with papillectomy. As such, thermal ablation should play a “bit” part in the treatment of ampullary adenomas, usually to “touch up” the margins of a resection site or to treat the adenomatous tissue (micro-adenomas) at the papillectomy site that frequently occur in Gardner’s syndrome.
Han and Kim have published a wonderful review of papillectomy. Their review included 800 cases published as case reports, abstracts and clinical series. Approximately one-quarter of patients had Gardner’s syndrome and three-quarters had sporadic adenomas. Patients presented with jaundice, cholangitis, pancreatitis, weight loss, or chronic cholangitis; or more commonly, were found during screening endoscopies. Lesions larger than 3 cm, as well as patients with Gardner’s adenomas, were less likely to have long-term successful treatment, although 80% of total patients had a long-term cure. Up to 10% of patients in some of the series were found to have underlying malignancy, requiring definitive surgery. Procedural complications approximated 10%, most commonly bleeding or pancreatitis, and the perforation rate ranged between 0% and 3%. There was a single procedural mortality in this composite of patient series.
Not only do patients with ampullary adenomas resected endoscopically need stent retrieval and papillectomy site inspection and biopsy at 4 to 6 weeks, they need long-term follow-up. Follow-up includes a baseline colonoscopy to insure that there are not concomitant colon adenomas, as well as evaluation of the ampullectomy site with a side-viewing scope. Contingent upon the patient and the assumption of complete eradication of the ampullary adenoma, screening should occur at least yearly for the first 3 years, and more frequently in patients with underlying genetic disorders.
In the review by Han and Kim, a 15% recurrence rate was noted. This rate is an estimate at best, as “recurrences” between 1 and 3 months postresection are much more likely related to incomplete resection than a true recurrence. Nevertheless, residual or recurrent disease is usually readily handled endoscopically, and the recurrence rates are probably comparable with the 12.5% rates reported in patients undergoing open surgical transduodenal papillectomy for ampullary adenomas.
Although it is clear that endoscopic papillectomy has supplanted surgery in most patients with amenable lesions, it is also clear that there are lesions that are better handled surgically. They include ulcerated or infiltrated lesions, patients who are found to have underlying malignancy in their resected specimens, Gardner’s patients in whom the papilla may be only one of hundreds of large and enlarging C-loop adenomas, lesions extending a significant distance into the PB tree, and those in whom the adenoma encompasses a considerable circumferential component of the C-loop wall. Figures 1 through 4 depict papillectomy in a high-risk patient who had significant dysplasia but no definite malignancy in his resected specimen.
Leung J, Lo SK, eds. Curbside Consultation in Endoscopy: 49 Clinical Questions, Second Edition (pp 183-187) © 2014 SLACK Incorporated.
- Catalano MF, et al. Gastrointest Endosc. 2004;59:225-232.
- Han J, Kim MH. Gastrointest Endosc. 2006;63:292-301.
- Heinzow HS, et al. Acta Gastroenterol Belg. 2011;74:509-515.
- Irani S, et al. Gastrointest Endosc. 2009;70:923-932.
- Patel R, et al. J Clin Gastroenterol. 2012;46:8-15.
- Winter JM, et al. J Gastrointest Surg. 2010;14:379-387.
- For more information on the Curbside Consultation in Gastroenterology series:
- Visit Healio.com/Books/Gastroenterology.