Transduodenal Resection of Peri-Ampullary Lesions

Elijah Dixon, M.D.,1 Charles M. Vollmer Jr., M.D.,2 Ajay Sahajpal, M.D.,2 Mark S. Cattral, M.D. M.Sc.,2

David R. Grant, M.D.,2 Bryce R. Taylor, M.D.,2 Bernard Langer, MD.,2 Steven Gallinger, M.D. M.Sc.,2

Paul D. Greig, M.D.2

1Department of Surgery, University of Calgary, Calgary, Canada2Department of Surgery, Toronto General Hospital, NU10-145, 421 University Avenue, M5G 2C4, Toronto, Ontario, Canada

Published Online: April 21, 2005

Abstract. Transduodenal resection (TDR) of lesions near the ampulla ofVater is an alternative to the Whipple pancreaticoduodenectomy. A ret-rospective analysis was performed to determine the long-term outcomeand the utility of intraoperative frozen section examinations in aidingoperative decision making in patients undergoing TDR. From 1992 to2002, 19 patients with an average age of 64.2 years (range: 33–84 years)underwent a transduodenal resection of a peri-ampullary lesion; medianfollow-up was 47 months (range: 2–100 months). Pathology of the lesionswas as follows: 11 with benign ampullary adenomas, including 4 withfamilial adenomatous polyposis (FAP); 7 with peri-ampullary adenocar-cinomas; and 1 with a benign stricture. Survival for the entire cohort is100%. In 12 cases an intraoperative frozen section was performed. Thespecificity and positive predictive value of the intraoperative histologywere both 100%, and the sensitivity and negative predictive value were57% and 38%, respectively. Three of the 4 patients with FAP haverecurrent adenomatous change; 2 of the 7 with carcinoma have metastaticadenocarcinoma. Transduodenal resection of peri-ampullary lesions ap-pears to be a safe alternative to radical resection for benign adenomasand selected carcinoma. Intraoperative frozen section assessment is rec-ommended in cases of potential adenocarcinoma.

Despite the description of transduodenal resection (TDR) forampullary lesions over a century ago (1899) by Halsted, itsutility in current practice is still controversial [1]. The adequacyof the resection as evidenced by disease-free survival and inci-dence of local recurrence may be inferior to the more extensive,yet higher risk, pancreaticoduodenectomy (PD). Transduodenalresection for small, benign, lesions has been shown to be a safealternative with low rates of recurrence [2], and it has beenrecommended for benign lesions less than 3 cm in size withouthigh-grade dysplasia. However, the local recurrence rate of be-nign villous tumors may be high: 43% at 10 years [3], including23% as new adenocarcinomas. Transduodenal resection of theampulla is safe, and may be considered in patients with earlystage cancers if both an R0 resection and a lymph node dis-

section are performed [4]. It may also be an acceptable option inpatients with high-risk medical comorbidities. The long-termoutcome of TDR for carcinoma of the ampulla of Vater hasbeen associated with a significant risk of late local recurrence.Although the morbidity and mortality associated with pancrea-ticoduodenectomy has decreased in the past two decades, mostseries show that local resections of the ampulla have comparablylower morbidity and mortality [2, 5, 6].

In patients with familial adenomatous polyposis (FAP) themanagement of duodenal and ampullary lesions is particularlycontroversial because of local recurrence. Some have suggestedthat repeated endoscopic therapy is a reasonable option for his-tologically advanced, non-cancerous lesions [7], whereas othershave recommended early radical surgery [8]. The relative roles ofendoscopy, local resection, and radical resection in the manage-ment of peri-ampullary lesions for FAP is still not clearly defined[9], and peri-ampullary carcinoma is now the leading cause ofdeath in patients afflicted with this syndrome. We retrospectivelyreviewed our results to determine what the long-term outcome isfor this heterogeneous group of patients.

Material and Methods

This is a retrospective analysis of all patients undergoing TDR ofthe ampulla of Vater between 1992 and 2002. Approval was ob-tained from the respective institutional review boards of theUniversity Health Network and the Mount Sinai Hospital, Uni-versity of Toronto, to review the charts of identified patients. Therecords and patient databases of six experienced (as a group,approximately 100 Whipple procedures per year) hepatobiliarysurgeons were reviewed. With the use of procedure codes forduodenal resection and transduodenal ampullectomy, 19 patientswere identified who underwent a local surgical resection of a peri-ampullary lesion. Using a standardized database collection form,a retrospective review was performed of in-patient hospital chartsas well as clinic charts. Data collected included age, sex, preop-erative bilirubin, presenting signs and symptoms, preoperativeendoscopy reports, operative procedure, blood loss, postoperativecomplications, pathology reports, and long-term survival and

Oral presentation at AHPBA meeting, Miami, Florida, USA, 2003..

Correspondence to: Paul D. Greig, M.D., Professor of Surgery, Uni-versity of Toronto, Toronto General Hospital, University Avenue, Tor-onto, Ontario, Canada, M5G 2C4, e-mail: paul.greig@uhn.on.ca

World J. Surg. 29, 649–652 (2005)

DOI: 10.1007/s00268-005-7578-6

follow-up information. In all cases the diagnosis was confirmed byreview of original pathology reports.

The operative procedure involved a longitudinal duodenotomyalong the second portion of the duodenum, opposite to the am-pulla of Vater. In some cases the gallbladder was removed as theinitial operative step, at which time a fine biliary sound was passeddown the common bile duct through the ampulla, thus helping toidentify the optimal location for the duodenotomy. Once identi-fied, a full-thickness excision of the ampulla was undertaken,incising through the distal bile duct and pancreatic duct. Prior tocomplete amputation of the two ducts, sutures were placed ineach with fine absorbable suture (usually 5-0 PDS) to prevent theducts from retracting out of view. The two ducts were thenapproximated along their medial borders with fine absorbablesutures. The combined orifice was then sutured circumferentiallyto the duodenum using full-thickness interrupted fine absorbablesutures. The duodenotomy was closed longitudinally, most com-monly in a single layer of interrupted sutures. Formal nodal dis-section was not performed. The technique has been describedpreviously [10].

Local recurrence was confirmed by endoscopy with biopsy.Similarly, metastatic disease was confirmed by computed tomog-raphy guided biopsy. None of the 19 patients had been lost tofollow-up at the time of analysis.

Descriptive statistics including means, medians, and ranges arereported. Statistical analysis was performed using SAS software(SAS Institute Inc., Cary, NC).

Results

Patient Characteristics (Tables 1 and 2)

The average age of these patients was 64.2 years (range: 33–84years). Ten patients were male and nine were female. Eighteenpatients were Caucasian and one patient was Asian. The mostcommon presenting symptoms were jaundice and pain/dyspepsia,followed by pancreatitis and weight loss, and radiographic iden-tification of the �double duct� sign. One patient presented withfecal occult blood positive stools. The average size of the lesionswas 2.7 cm (range: 1–10 cm) as determined by final pathology.Follow-up of all 19 patients is complete, with the average follow-up period being 45 months (range: 2–100 months). Four patientswere known to have FAP syndrome (Tables 1 and 2).

Operative Characteristics

Nineteen patients underwent a TDR of the Ampulla of Vater forbenign (Table 1) or malignant (Table 2) lesions. Of these, twopatients had the TDR converted to a pancreaticoduodenectomy(PD) based on identification of margin-positive adenocarcinomaidentified by frozen section. Three were extended TDRs; in twocases resection of the duodenal wall up to the first part of theduodenum was performed, and in the third a circumferentialresection of the second portion of the duodenum was performed.Average intraoperative blood loss was estimated at 230 cc (range:50–450 cc).

Table 1. Benign lesions

Patient Age Sex FAP Frozensection

Procedure Post-operativediagnosis

Dysplasia Follow-up(mo)

R0 Resection Recurrence(mo)

1 59 F Yes Yes TDA VA Mild–Mod. 86 No Radial margin + Local VA 10 mo2 68 F Yes No TDA VA High grade 82 No. distal CBD + No3 60 M Yes Yes TDA TA Mild 100 Yes Local VA 12 mo4 45 F Yes No TDA VA High grade 60 No distal CBD + Local VA 46 mo5 72 F No Yes TDA VA Mild 14 Yes No6 73 M No No TDA VA Mild 41 Yes No7 80 F No No TDA Benign stricture NA 71 NA No8 33 M No Yes TDA VA Mild 3 Yes No9 62 F No Yes TDA VA Mild 28 No No10 37 M No No TDA VA Mild 68 Yes No11 68 M No No TDA VA Mild 92 Yes No12 56 F No No TDA VA Mild 2 No positive duod. Marg. No

M: male; F: female; FAP: familial adenomatous polyposis TDA: transduodenal ampullectomy; VA: villous adenoma; TA: tubular adenoma; mod.:moderate; CBD: common bile duct; NA: not applicable; duod. Marg.: duodenal margin.

Table 2. Malignant lesions

Patient Age Sex Frozen section Procedure Post-operative diagnosis Follow-up (mo) R0 Resection Recurrence (mo)

1 67 M Yes–VA TDA T1 papillary cancer Ampulla 44 Yes Local pane,and Liver mets 36 mo

2 84 M Yes–VA TDA T1 AdenoCA Ampulla 9 Yes No3 76 M Yes–VA TDA T1 AdenoCA Ampulla 67 Yes No4 72 M Yes–cancer margin

negativeTDA T1 Adenoca Ampulla 41 No–pancreatic

margin positiveNo

5 81 M Yes–AdenoCApreviouslysuspected/known

TDA T1 AdenoCA Ampulla 28 Yes Yes–bilobar livermets 23 mo

6 72 F Yes-AdenoCA TDA fi T1 AdenoCA Ampulla 8 Yes No7 55 F Yes-AdenoCA TDA fi T2N1 2/5 lymph nodes

+ AdenoCA Ampulla30 Yes No

PD: pancreaticoduodenectomy; AdenoCA: adenocarcinoma; panc. – pancreatic; mets – metastases.

650 World J. Surg. Vol. 29, No. 5, May 2005

Outcomes

To date overall survival is 100% for patients with benign (Table 1)and malignant (Table 2) lesions. Of the 11 patients with benignadenomatous lesions, 3 have had a local recurrence (Table 1). All3 were patients with FAP. These local recurrences occurred at10,12, and 46 months. In the 7 patients with malignant disease, 2have had recurrence. One patient had a combined local anddistant recurrence in the uncinate process of the pancreas and theliver, respectively. This occurred at 36 months. In the secondpatient, bilobar liver metastases were detected at 23 months.

Frozen Section Analysis

Intraoperative frozen section (FS) analysis was used selectivelybased on the intraoperative assessment of the operating surgeon.Twelve patients had intraoperative frozen sections performed.Three were done to assess the margins of resection (patients 4, 5,and 7), and the other 9 were done to determine if malignantdisease was present. Margins analysis was correct in 2 of 3 pa-tients (patient 4 had a positive pancreatic resection margin). Inthe 9 frozen section evaluated for cancer, 6 were correct (5 be-nign, 1 known malignant) and 3 were incorrect (all false nega-tives). In five cases the frozen section affected intraoperativedecision making: three patients had positive resection margins re-resected, and two procedures were converted to PD for malig-nancy. Overall, the frozen section examinations correctly agreedwith the final pathology in 75% of the cases. However, in threecases the frozen section examination did not detect cancers foundon final pathology (patients 1, 2, 3, Table 2). The outcomes of thefrozen section examinations in the detection of cancer are asfollows: sensitivity—57%, specificity—100%, positive predictivevalue—100%, and negative predictive value—62% (Table 3).

Complications

One patient developed an early postoperative biliary fistula thatrequired laparotomy and drainage. One patient required percu-taneous drainage of an intrabdominal abscess. One patient had ananastomotic leak resulting in a combined biliary and pancreaticfistula that resolved with conservative management. One patientdeveloped a chronic incisional pain syndrome. Two patientsdeveloped incisional hernias, which were repaired. Delayed gas-tric emptying resulted in a prolonged admission for one patient,managed with intravenous nutrition.

Discussion

The indications for local resection of the ampulla versus pan-creaticoduodenectomy are not widely agreed upon. A recent studyhas recommended PD for most villous tumors of the ampullabecause of the inaccuracy of preoperative biopsy, which can have a

false negative rate of 25%–56% [11]. Others have recommendedlocal resection with intraoperative frozen section for small benignlesions [2]. The morbidity and mortality associated with PD ascompared to local ampullary resections is high [2, 5, 6].

This review illustrates three important points. The first is theutility of intraoperative frozen section (FS). Intraoperative FSanalysis was performed selectively based on the intraoperativeassessment of the operating surgeon, each of whom had extensiveexperience in hepatobiliary/pancreatic surgical diseases. (All casesthat did not have FS examination performed had benign pathol-ogy consistent with the operating surgeon�s assessment). Al-though the specificity and positive predictive value of FS werehigh, the sensitivity and negative predictive value were very low. Itmay be that with closer examinations (more sections) the numberof false negatives could be reduced. Despite these limitations,given the high frequency with which FS examination altered in-traoperative decision making, we recommend that it should bemade an integral part of a TDR of the ampulla for adenomatousdisease. To improve the results of FS, a number of sections shouldbe taken from each margin.

The second point regards the management of early stage peri-ampullary carcinoma. Of the 5 patients who underwent TDR formalignancy, each had a complete R0 resection and all are alive,with follow-up exceeding 2 years in 4 cases. Two have hadrecurrence: both with hepatic metastases and only one withrecurrence in the uncinate process of the pancreas. These resultsin selected patients support the use of TDR in the treatment ofearly stage, low-risk cancers [4]. In general, however, we supportthe recommendation that PD is the procedure of choice except inthose patients with prohibitive medical comorbidities [3, 6], pro-vided the operation is performed by a team with a documentedrecord of low morbidity and mortality. In patients with significantcomorbid medical disease the use of TDR is a reasonable alter-native to manage the local disease.

Third, patients with FAP and villous tumors of the ampulla areat high risk of local recurrence after TDR; three of the four pa-tients in our series had a local recurrence. This is similar to seriesin which 7 of 11 patients treated with TDR had recurrence [3]. Inour series, none of the recurrences have been malignant, andthese patients continue to be surveilled. In the future, PD orpancreas-sparing duodenectomy remain an option. We recom-mend that benign villous adenomas in patients with FAP betreated with TDR and regular endoscopic follow-up to detectrecurrences. However, recognizing the high likelihood of adeno-matous recurrence, and the fact that up to 12% of patients withFAP die of peri-ampullary cancers [9], the PD or pancreas-sparing duodenectomy remains an option, especially in patientswith extensive villous adenomas throughout the duodenum orthose who are young.

The limitations of this study are its retrospective nature and therelatively small number of patients. However, this is one of thelarger series of TDR reported to date. Moreover, infrequency ofampullary carcinoma and the lack of therapeutic equipoiseregarding its management make it very unlikely that a prospectiverandomized trial will ever be attempted. It is hoped that theinformation in this analysis, in the context of previous reports, isvaluable in determining therapeutic management for these pa-tients in the future.

Transduodenal resection of peri-ampullary lesions appears tobe a safe alternative to radical resection for benign adenomas and

Table 3. Results of frozen section biopsy.

Final/Pathology

Frozen sectionpathology

Adenocarcinoma Noadenocarcinoma

Adenocarcinoma 4 0No adenocarcinoma 3 5

Dixon et al.: Transduodenal Ampullectomy 651

selected carcinoma. Intraoperative frozen section assessment isrecommended to assess resection margins, and in cases of po-tential adenocarcinoma.

References

1. Halsted W. Contributions to the surgery of the bile passages, espe-cially of the common bile duct. Boston Med. Surg. J. 1899;141

2. Clary B, Tyler D, Dematos P, et al. Local ampullary resection withcareful intraoperative frozen section evaluation for presumed benignampullary neoplasms. Surgery 2000;127:628–633

3. Farnell MB, Sakorafas GH, Sarr MG, et al. Villous tumours of theduodenum: reappraisal of local vs. extended resection. J. Gastrointest.Surg. 2000;4:13–23

4. Beger HG, Treitschke F, Gansauge F, et al. Tumor of the ampulla ofVater: experience with local or radical resection in 171 consecutivelytreated patients. Arch. Surg. 1999;134:526–532

5. Rattner D, Fernandez-del Castillo C, Brugge W, et al. Defining thecriteria for local resection of ampullary neoplasms. Arch. Surg.1996;131:36–37

6. BranumG, Pappas T,Meyers W, et al. The management of tumours ofthe ampulla of Vater by local resection. Ann. Surg. 1996;224:621–627

7. Alarcon FJ, Burke CA, Church JM, et al. Familial adenomatouspolyposis: efficacy of endoscopic and surgical treatment for advancedduodenal adenomas. Dis. Colon Rectum 1999;42:1533–1536

8. Wallace MH, Phillips RK. Preventative strategies for peri-ampullarytumours in FAP. Ann Oncol. 1999;10:(Suppl 4):201–203

9. Belchetz LA, Berk T, Bapat BV, et al. Changing causes of mortality inpatients with familial adenomatous polyposis. Dis. Colon Rectum1996;39:384–387

10. Blumgart LH, Fong Y. Surgery of the liver and biliary tract 3. Lon-don: W.B. Saunders, 2000, pp 1090–1099

11. Jordan PH, Ayala G, Rosenberg WR, et al. Treatment of ampullaryvillous adenomas that may harbor carcinoma. J. Gastrointest. Surg.2002;6:770–775

652 World J. Surg. Vol. 29, No. 5, May 2005