Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Papillectomy and Ampullectomy
Ampullary neoplasms (see Chapter 38 ) are rare, with an annual incidence of 3000 in the United States and a reported prevalence of 0.04% to 0.12% in autopsy series. Endoscopically the papilla can appear enlarged and abnormal because of various tumors (benign or malignant) and other etiologies such as papillitis, gastric foveolar metaplasia, or pancreatic acinar hyperplasia ( Fig. 25.1 ). Ampullary tumors can be classified based on their layer of origin: those that arise from the epithelium (e.g., adenomas, adenocarcinomas, lymphomas) and those that arise from the subepithelium (e.g., neuroendocrine tumors, lipomas, leiomyomas, gastrointestinal stromal tumors, lymphangiomas, fibromas, hamartomas). Ampullary adenomas are by far the most common of these. Although uncommon in the general population (representing <10% of all periampullary tumors), they are increased twofold to threefold in genetic polyposis syndromes, especially familial adenomatous polyposis (FAP) and its variants. Between 40% and 100% of FAP patients will also develop duodenal adenomas, which are frequently numerous with a similar malignant potential. Ampullary adenomas can arise from the surface or the inner epithelium of the ampulla.
Enlarged prominent papilla that could be mistaken for an ampullary adenoma because of papillitis ( A ), gastric foveolar metaplasia ( B ), and pancreatic acinar hyperplasia ( C ).
Historically, ampullary adenomas presented late, with a high incidence of underlying malignancy. Endoscopic management in the early years consisted mainly of palliation of obstructive jaundice with biliary sphincterotomy and stent placement. Patients often become symptomatic when the lesions are large enough to cause obstruction, presenting with cholestasis, cholangitis, pancreatitis, nonspecific abdominal pain, and, less commonly, bleeding. With increased awareness among endoscopists and the increasing use of cross-sectional imaging (computed tomography [CT] and magnetic resonance imaging [MRI]), these lesions are being recognized at earlier stages with a lower incidence of underlying malignancy. Asymptomatic lesions are being recognized more commonly in patients undergoing endoscopy for unrelated reasons such as gastroesophageal reflux disease (GERD), surveillance of Barrett esophagus, and dyspepsia or, alternatively, in patients undergoing surveillance for FAP.
Treatment Options
Once diagnosed with a biopsy-proven ampullary adenoma, patients have four options: close endoscopic surveillance with biliary stent placement (for cholestasis) or without biliary stent placement endoscopic papillectomy, surgical ampullectomy/local resection, or pancreaticoduodenectomy (Whipple operation). There are no clinical trials comparing one approach to another. The decision regarding treatment choice is influenced by various factors, including patient preference, associated comorbidities that may affect fitness for surgery, local endoscopic and surgical expertise, characteristics of the lesion, and whether sporadic or in the setting of FAP. For patients with large lesions (greater than 4 to 5 cm), presence of high-grade dysplasia or carcinoma in situ, or obvious nodal disease on cross-sectional imaging or endoscopic ultrasonography (EUS), pancreaticoduodenectomy should be considered. However, small series have demonstrated the ability to completely resect focal, low-grade, early-stage T1 ampullary adenocarcinoma via endoscopic papillectomy.
Although associated with the highest cure rates and lowest recurrence rates, pancreaticoduodenectomy carries a morbidity rate of 25% to 63%, with mortality ranging from 0% to 13% (higher rates are reported in patients with malignant disease). Surgical local resection is associated with lower morbidity (14% to 27%) and mortality (0% to 4%), but recurrence rates vary from 0% to 32% (mean, 18% to 20%), thus necessitating continued postoperative endoscopic surveillance. In a retrospective review of 38 patients, surgical ampullectomy was associated with lower operative time (169 minutes vs 268 minutes), estimated blood loss (192 mL vs 727 mL), mean hospital stay (10 days vs 25 days), and overall morbidity (29% vs 78%) compared with pancreaticoduodenectomy. As colonoscopic polypectomy has become routine for even large colonic polyps, and with the acceptable morbidity, mortality, and recurrence rates of surgical local resection for papillary adenomas, endoscopic papillectomy emerged as an acceptable alternative to surgery.
Endoscopic papillectomy was first described by Suzuki et al. in 1983, and the first large case series was reported by Binmoeller et al. in 1993. Over the last 2 decades, many more studies have been published showing high success rates, low morbidity, and minimal mortality. Endoscopic papillectomy has therefore gained increasing acceptance as the treatment of choice for the vast majority of patients with ampullary adenomas.
Considerations in FAP
In patients with FAP, ampullary adenomas develop in 80% in their lifetime, with up to a 5% risk of developing malignancy. The management of ampullary adenomas in the setting of FAP is further complicated because complete excision does not eliminate the risk of recurrence or new upper gastrointestinal tract cancers. Fortunately, the risk of histologic progression of upper gastrointestinal tract adenomas in patients with FAP appears to be low. In one of the largest studies on surveillance of duodenal and ampullary adenomas in 114 patients with FAP, duodenal polyps progressed in size in 26%, number in 32%, and histology in 11% of patients. Morphology and histology of the papilla progressed in 14% and 11% of patients, respectively, with only one patient developing a periampullary cancer over a mean follow-up of 51 months. This has prompted some experts to propose surveillance endoscopy with biopsies only, rather than excision, for FAP patients with ampullary adenomas in the absence of rapid growth or high-grade dysplasia on endoscopic biopsy. The obvious benefit of this approach is avoidance of the risks associated with excision of these lesions. The limitation of surveillance with biopsies alone is the potential to miss occult foci of high-grade dysplasia or carcinoma. In a study of 33 patients with ampullary adenomas (not necessarily FAP) in whom endoscopic biopsies were obtained, carcinoma was found at final surgical pathology in 5/10 patients with high-grade dysplasia and in 3/19 patients with low-grade dysplasia. Because of these competing issues, the management of ampullary adenomas in FAP patients is determined on a case-by-case basis, taking into consideration all of the aforementioned factors in addition to the patient's wishes and risk tolerance.
Technique ( Box 25.1 )
The role of endoscopic papillectomy was reviewed in a guideline statement published by the American Society for Gastrointestinal Endoscopy in 2006.
Box 25.1
Summary of the Technique
-
Staging of ampullary adenomas with cross-sectional imaging (computed tomography or magnetic resonance imaging) is not routine, although in the setting of significant weight loss and dilated ducts should be done to evaluate for the presence of nodal and metastatic disease.
-
Endoscopic ultrasonography is very helpful in defining lesion size, depth of invasion, ductal involvement, and anatomy but is not routine, especially in small lesions and in the absence of jaundice, history of acute pancreatitis, or dilation of the pancreatic or bile ducts.
-
Endoscopic retrograde cholangiopancreatography before papillectomy is indicated in all patients to assess whether intraductal involvement is present and to identify pancreatic duct anatomy for conditions such as pancreas divisum.
-
A duodenoscope is used to perform the papillectomy in a similar manner to colonoscopic snare polypectomy, usually using blended current.
-
Thermal ablation may be used in select cases as adjunctive therapy.
-
Prophylactic pancreatic sphincterotomy and stent placement should be attempted in all patients.
-
Routine biliary sphincterotomy is recommended, although biliary stent placement is not routine in the presence of good biliary drainage.
-
Surveillance is critical after complete resection because of known local recurrences.
Initial Endoscopic Assessment
Conventional Endoscopy
Visualization of the ampulla is possible with a conventional forward-viewing endoscope but is best performed with a duodenoscope. A large periampullary diverticulum can hinder visualization ( Fig. 25.2 ). Ampullary adenomas can have varied appearances, ranging from appearing normal in the setting of FAP, to a slight enlargement, to flat laterally spreading, to granular or polypoid, with or without ulceration ( Fig. 25.3 ). The use of endoscopic tools such as narrow band imaging (NBI), flexible spectral imaging color enhancement (FICE), or magnification endoscopy may help further characterize lesions in certain cases. Although some endoscopic features alone should raise the suspicion of an associated underlying malignancy (large, friable, ulcerated, indurated lesions), endoscopic biopsies are needed to confirm the pathology. Endoscopic biopsies, however, can be suboptimal at detecting underlying malignant foci. The reported rates of malignancy in ampullary adenomas are approximately 20% to 30%. Detection rates for cancer present within adenomas based on biopsy alone range from a dismal 40% to 89%. These arguments may suggest favoring a radical surgical approach in all patients; however, this is mostly based on data obtained from surgical series. Nowadays, many ampullary lesions are diagnosed at routine endoscopy, with a much higher percentage being detected at earlier stages. In the more recent and larger endoscopic series (over 100 patients in each), incidental or asymptomatic presentation of ampullary adenomatous lesions was seen in 25% to 33%. Rates of malignancy in papillectomy specimens ranged from 6% to 8%. In addition, there are clinical, endoscopic, and imaging features that should give the clinician pause before considering endoscopic papillectomy, even if biopsies alone do not reveal cancer—specifically, clinical features such as significant weight loss and jaundice and endoscopic features such as large (>4 cm), friable, ulcerated, fixed lesions and evidence of dilated ducts, especially with intraductal extension, as seen on EUS, MRCP, or during cholangiography at the time of removal. In our series of 102 patients who underwent papillectomy, 8 (7%) were found to have invasive cancer. Of these eight patients, two were referred by surgeons because of significant comorbidities precluding pancreaticoduodenectomy. Thus only 6 of 102 (5%) patients who appeared endoscopically resectable and who underwent papillectomy had invasive cancer on final pathology. Therefore, although it is true that biopsy alone has a definable miss rate in diagnosing invasive cancer, only a minority of patients with invasive cancer seem to have endoscopically resectable lesions. Most authors propose that EUS in selected patients could help decrease this unnecessary papillectomy rate even further, with many centers performing EUS routinely for all ampullary adenomas. More recently, in a study of 56 ampullary tumors, histology and immunohistochemical characteristics helped further with selecting the appropriate surgical or endoscopic approach. They demonstrated that periampullary tumors, intestinal histology, and high CK20-positive rate can be regarded as good indications for endoscopic papillectomy. On the other hand, tumors that are either pancreatobiliary or intestinal type without CDX2 expression have a higher chance of intraductal involvement and may be better suited for a surgical approach.
Endoscopic Ultrasonography and Intraductal Endoscopic Ultrasonography
Some authors recommend performing EUS in all patients with ampullary adenomas, whereas others recommend EUS only for selected cases. As mentioned above, clinical features of weight loss and jaundice, cross-sectional imaging features of dilated ducts, and endoscopic appearance (friable, ulcerated, indurated, and in our practice lesions >2 cm) should prompt performance of EUS before papillectomy. EUS is very useful in evaluating tumor size, depth of invasion of the duodenal wall, and involvement of the biliary or pancreatic ducts and periampullary lymph nodes ( Fig. 25.4 ). The T-staging accuracy of EUS is very good (83% to 90%) and superior to CT and MRI. N-staging accuracy in some studies is reported to be better than CT (84% vs 68%), whereas in others it is reported to be comparable to CT and MRI (68% vs 59% vs 77%). Prior sphincterotomy and the presence of a biliary or pancreatic stent, however, can decrease EUS T-stage accuracy.
The optimal management of patients with intramucosal cancer (pT1) remains debatable. In these patients, lymphatic or vascular invasion and lymph node metastases do not occur. Local resection can be justified if biliary or pancreatic duct infiltration can be confidently excluded. The evaluation of tumor invasion of the sphincter of Oddi remains challenging. Intraductal endoscopic ultrasonography (IDUS) can be performed in a transpapillary or percutaneous fashion. IDUS has excellent resolution because of the use of high-frequency ultrasound (20 to 30 MHz) compared with standard EUS (at most usually 7.5 to 10 MHz). There are three studies comparing EUS with IDUS in ampullary cancer. Itoh et al. reported the diagnostic accuracy of IDUS and EUS in 32 patients with ampullary cancer who underwent surgical resection. The TNM staging accuracy of IDUS was comparable to EUS (88% vs 90%) but the depiction of the sphincter of Oddi and detailed staging were better with IDUS. Menzel et al. reported the results of a prospective study on 27 patients with ampullary neoplasm (adenoma in 12, adenocarcinoma in 15) who underwent surgical management. They concluded that IDUS was significantly superior to EUS with regard to tumor visualization and staging (staging accuracy 93% vs 62%). Ito et al. performed EUS and transpapillary IDUS in 40 patients (adenocarcinoma in 33, adenoma in 7). IDUS had a slightly better T-stage accuracy (78% vs 62%) but was no better at detecting ductal involvement (90% vs 88%). Because of the paucity of data and similar performance of IDUS and EUS, IDUS cannot be routinely recommended at this time. However, as probe technology improves it may become more commonly used to differentiate more subtle, early invasive cancers of the ampulla, thus precluding unnecessary papillectomies.
Endoscopic Retrograde Cholangiopancreatography
Although EUS is generally performed before resection to improve staging accuracy, endoscopic retrograde cholangiopancreatography (ERCP) is performed at the time of papillectomy to exclude intraductal tumor involvement and to perform pancreaticobiliary sphincterotomy and stent placement. If EUS is unavailable or if the findings at EUS are equivocal, performance of ERCP is critical before resection ( Fig. 25.5 ). Surgical referral should be considered if there is evidence of intraductal involvement, especially if there is extension >1 cm into the biliary or pancreatic duct. Cholangiopancreatography outlines the ductal anatomy to help access the pancreatic and bile ducts after papillectomy. In addition, the presence of pancreas divisum can obviate the need for pancreatic duct stent placement.
Endoscopic Papillectomy
Endoscopic resection is limited to the mucosa and submucosa of the duodenal wall and the tissue around the biliary and pancreatic duct orifices located at the major duodenal papilla. Endoscopic papillectomy makes it difficult to remove tumor tissue invading the common bile duct or pancreatic duct. In clinical practice, the terms endoscopic papillectomy and endoscopic ampullectomy are used interchangeably. However, ampullectomy consists of circumferential resection of the ampulla of Vater, with separate reinsertion of the bile duct and pancreatic duct into the duodenal wall. This cannot be achieved using translumenal endoscopic therapy and necessitates a surgical duodenotomy and resection of pancreatic tissue in the area of anatomic attachment of the ampulla to the duodenal wall. Therefore the term endoscopic papillectomy is more appropriate than endoscopic ampullectomy.
The goal of endoscopic resection is to achieve complete excision of the ampullary neoplasm. There are currently many different techniques to perform a papillectomy, with no consensus on one method and a lack of large comparative studies. In deciding how best to remove an ampullary adenoma, the endoscopist must clearly define the margins of the lesion (which may sometimes be difficult in very flat, spreading lesions or when associated with a periampullary diverticulum). In general, papillectomy is accomplished in a manner similar to colonoscopic snare polypectomy using a duodenoscope. Although there are no data to support administration of antibiotics before endoscopic papillectomy, the procedure is similar to an ERCP and other submucosal resections. We routinely give antibiotics before papillectomy and in cases of concern for injury to the duodenal muscularis propria, and we continue for 2 to 3 days after resection.
You're Reading a Preview
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here