Clinical outcomes of endoscopic resection for nonampullary duodenal high-grade dysplasia and intramucosal carcinoma

• Introduction
• Case series
• Patients and methods
• Results
• Discussion
• References

This study retrospectively analyzed the clinical outcomes of endoscopic resection of 26 sporadic (i. e., not associated with polyposis syndrome) nonampullary duodenal lesions representing high-grade dysplasia or intramucosal carcinoma (duodenal HGD/IMC) in 23 patients. No severe complications such as perforation were observed, but three cases of delayed bleeding were seen. The use of endoscopic clips significantly decreased the delayed bleeding rate (0 /19, 0 %) compared with cases in which clips were not used (3 /7, 42.9 %; P = 0.013, χ² test). Eighteen lesions (69.2 %) were removed by en bloc resection. The follow-up period after resection was 25.5 ± 23.3 months. Two lesions (7.7 %) that recurred locally were detected at the first surveillance endoscopy 3 months after resection. These lesions were 22 and 15 mm in size respectively and were resected piecemeal. Endoscopic resection is an effective and safe procedure for treating duodenal HGD/IMC. En bloc resection and prophylactic clip usage are encouraged.

Introduction

Nonampullary duodenal adenoma is relatively common in patients with familial adenomatous polyposis, but sporadic nonampullary duodenal adenoma is rare. Endoscopic mucosal resection (EMR) of duodenal neoplasms was first performed in 1992 [1]. However, the duodenum is one of the most difficult areas for endoscopic resection of tumors [2], and for this reason there are few studies reporting the outcome of endoscopic resection of sporadic nonampullary duodenal adenoma, and even fewer outcome of endoscopic resection of carcinoma [2] [3] [4] [5] [6].

In the present study, we evaluated the safety and efficacy of endoscopic resection of sporadic nonampullary duodenal high-grade dysplasia and intramucosal carcinoma (duodenal HGD/IMC) by investigating outcomes, complications, and prognosis.

Case series

Patients and methods

Between January 2004 and June 2011, 69 patients with duodenal tumors were treated in the Department of Gastroenterology, Chiba University Hospital (patients with advanced duodenal cancer with metastasis or invasion to other organs were admitted directly to chemotherapy or the surgical department). Forty-three of the 69 patients had nonampullary tumors, and the other 26 had ampullary tumors. To evaluate the clinical outcomes of endoscopic resection of sporadic nonampullary duodenal tumors, we treated 37 of the 43 patients with nonampullary duodenal tumors by endoscopic resection. The other 6 patients were excluded from this treatment for the following reasons: two underwent surgical resection, one received chemotherapy, and of the remaining three, one was diagnosed with a gastrointestinal stromal tumor, one with a carcinoid lesion, and one with familial adenomatous polyposis with ampullary carcinoma.

Out of the 37 patients, we retrospectively analyzed the records of 23 with sporadic nonampullary duodenal adenomatous lesions with adenocarcinomatous components. All lesions were inspected using chromoendoscopy with 0.08 % indigo carmine spraying, and biopsied using a conventional gastroscope (Olympus Corp., Tokyo, Japan; Fujifilm Corp., Tokyo, Japan). The lesions were resected using the snare-assisted technique, where normal saline solution or a glycerol-based solution in combination with indigo carmine is injected into the submucosal layer with a disposable 23- or 25-gauge injection needle (Olympus). Most of the lesions were resected by the EMR method using an oval disposable braided snare with a diameter of 10 or 20 mm (SnareMaster; Olympus) or a normal-type snare (SD5L-1; Olympus). We resected some lesions by EMR (ESO) using a cap-fitted panendoscope [7] with a half-moon braided snare (SD-7P-1; Olympus), using the two-channel method (strip biopsy) [8], or by endoscopic submucosal dissection using a DualKnife (KD-650L; Olympus). Although complete en bloc resection was attempted for all lesions, it could not be performed in some cases because of insufficient tumor lifting, even by injection. Some post-resection ulcers were closed with endoscopic clips (Olympus) for prophylaxis or to treat procedure-related bleeding. Some cases of intraprocedural bleeding were treated with endoscopic clips, argon plasma coagulation, or a combination of both. All of the specimens were reviewed by specialists in gastrointestinal pathology and evaluated based on the analysis of the vertical and lateral margins. All lesions resected piecemeal were diagnosed as having an “unclear lateral margin.” We permitted the patients to begin a liquid diet 2 days after the resection, and patients were discharged within 6 days after resection. All patients received proton pump inhibitory therapy (omeprazole, 20 mg/day) for at least 2 weeks after resection. Follow-up endoscopy was performed at 3, 6, and 12 months after resection and annually thereafter. When recurrence was observed after resection, we stopped the follow-up and again performed endoscopic treatment. All statistical analyses were performed using the Dr. SPSS 2 statistical software package (SPSS Inc., IBM Japan, Tokyo, Japan).

Results

The mean age of the 23 patients with sporadic nonampullary duodenal adenomatous lesions with adenocarcinomatous components who underwent endoscopic resection was 64 ± 12 years (range, 39 – 82 years). Of the 23 patients, 18 were men and 5 were women. One patient had four lesions, but he did not show signs of polyposis disease. Thus, 26 lesions in 23 patients were resected. One patient was receiving aspirin and clopidogrel. This regimen was stopped 10 days before resection, the patient underwent heparinization, and no intraprocedural or delayed bleeding was observed.

Five of the 23 patients (21.7 %) had simultaneous or metachronous occurrence of gastric cancer, and all five of them were positive for Helicobacter pylori infection and had H. pylori-related gastritis with multifocal atrophy. The lesion characteristics are summarized in [Table 1]. Nineteen lesions (73.1 %) were located on the oral side of the major papilla. Fourteen lesions were located in the second duodenum, and five lesions were in the first duodenum. All lesions in the patients with H. pylori-related gastritis with multifocal atrophy were located on the oral side of the major duodenal papilla (11 /11, 100 %), representing a significantly higher prevalence than that in patients without H. pylori-related gastritis with multifocal atrophy (8 /15, 53.3 %; P = 0.01, χ² test).

All of the lesions were completely resected in a single session using a gastroscope, and 18 lesions (69.2 %) were removed en bloc. The mean size of the lesions resected piecemeal was 14.4 ± 5.4 mm (range, 6 – 22 mm), significantly larger than the mean size of the lesions removed en bloc (8.1 ± 3.8 mm; range, 3 – 18 mm; P = 0.008, Mann – Whitney test). Pathologically, 10 lesions proved to be intramucosal carcinoma without lymphovascular invasion and 16 lesions were high-grade dysplasia [9]. No lesion was classified as submucosal invasive carcinoma after endoscopic resection. In 5 of 26 lesions (19.2 %), tumor cells were detected on the lateral margin, and in 13 lesions (50.0 %), it was unclear whether tumor cells were present on the lateral margin. The mean size of the lesions resected with positive or unclear pathological findings on the lateral margin was 11.9 ± 5.1 mm (range, 5 – 22 mm), significantly larger than the mean size of the lesions with negative pathological findings on the lateral margin (6.6 ± 3.1 mm; range, 3 – 12 mm; P = 0.008, Mann – Whitney test).

Intraprocedural bleeding occurred in nine cases. Among lesions resected piecemeal, the incidence of intraprocedural bleeding was 75.0 % (6 /8), significantly higher than among lesions removed en bloc (3 /18, 16.7 %; P = 0.005, Fisher’s exact test). The mean size of the lesions with intraprocedural bleeding was significantly larger than that of the lesions without intraprocedural bleeding (13.2 ± 5.5 mm [range, 6 – 22 mm] vs. 8.4 ± 4.2 mm [range, 3 – 18 mm]; P = 0.029, Mann – Whitney test). In 19 lesions, endoscopic clips were used for prophylaxis or treatment of procedure-related bleeding ([Fig. 1]). Delayed bleeding occurred within 2 days after resection in three cases where endoscopic clips were not placed after resection. All of these bleeds were treated successfully with endoscopic clips. The delayed bleeding rate was significantly decreased by the use of endoscopic clips (0 /19, 0 %) compared with the cases in which clips were not used (3 /7, 42.9 %; P = 0.013, χ² test), whereas size, the treatment method (EMR, EMR using a cap-fitted panendoscope, strip biopsy, or ESD), pattern of resection (en bloc vs. piecemeal), and intraprocedural bleeding did not significantly affect the delayed bleeding rate. Blood transfusion was not required in any case, and perforation was not observed in any patient. Although 24 lesions (92.3 %) showed no recurrence for 27.4 ± 23.3 months (range, 3 – 79 months), the 2 lesions (7.7 %) that were 22 mm and 15 mm at the first treatment recurred locally, as was detected at the first surveillance endoscopy 3 months after resection. These two lesions were larger than the others (9.3 ± 4.6 mm) and had been resected piecemeal, and the lateral margins of the specimens were positive and unclear.

Discussion

This is the first report to analyze a large number of patients with duodenal HGD/IMC who were treated by endoscopic resection. Ten of 26 lesions were intramucosal carcinoma and the other 16 lesions were high-grade dysplasia. In accordance with Japanese criteria, all of these 26 lesions were diagnosed as carcinoma based on the concept of carcinoma in situ [9].

Previous studies have reported delayed bleeding rates of 0 % – 20 % after endoscopic resection of nonampullary duodenal adenomatous lesions [2] [3] [4] [5] [6] [7] [8]. In all cases of delayed bleeding in the present study endoscopic clips had not been placed after resection. Like the present authors, Lépilliez et al. found that the use of prophylactic endoscopic clips after the resection of duodenal tumors reduced the risk of delayed bleeding [4], in contrast to their deployment in colon tumors [10]. We suggest that this discrepancy results from the severe conditions prevailing in the duodenum. Because strong stimulants such as pancreatic juice and bile could interfere with the healing of a postprocedural ulcer in the duodenum, ulcer closure with an endoscopic clip is more important in the duodenum than in other organs. We performed follow-up endoscopy at 3, 6, and 12 months after treatment, regardless of whether the resection sites were closed with clips. There was no bleeding at the follow-up endoscopy, and therefore we were not concerned about delayed bleeding at over 3 months after resection, especially when the lesion was closed with clips.

In all cases where lesions recurred in the present study, the lesions had a diameter greater than or equal to 15 mm and had been resected piecemeal. In the present study, large tumor size was a risk factor for piecemeal resection. We determined that en bloc resection is important in the endoscopic resection of duodenal HGD/IMC.

In Western countries, primary duodenal cancers are more frequently located on the anal side of the papilla than on the oral side [11], and exposure to bile is considered to be a factor in carcinogenesis [12]. In contrast to the West, in Japan primary duodenal cancer occurs with a high frequency on the oral side of the papilla [13], as observed in the present study. We speculate that this discrepancy may result from the high H. pylori infection rate in Japanese people. Helicobacter pylori infection affects the acidic environment of the duodenum, and this change in acid exposure, especially on the oral side of the papilla, affects the duodenal mucosa. In the present study, the lesions in the patients with H. pylori-related gastritis with multifocal atrophy were indeed significantly more often located on the oral side of the major duodenal papilla than those in patients without H. pylori-related gastritis. However, we did not clarify changes to the duodenal mucosa, such as gastric metaplasia or duodenitis, by pathological analysis. Moreover, compared to the high rate of H. pylori infection in Japan, only a small number of people develop duodenal carcinoma. Further analysis is therefore necessary to determine the association between gastric cancer, H. pylori infection, and duodenal HGD/IMC.

In conclusion, we have demonstrated that endoscopic resection is a safe and effective procedure for treating duodenal HGD/IMC. To prevent local recurrence of duodenal HGD/IMC and delayed bleeding, en bloc resection and prophylactic use of endoscopic clips are encouraged.

Competing interests: None

• References

• 1 Obata S, Suenaga M, Araki K et al. Use of strip biopsy in a case of early duodenal cancer. Endoscopy 1992; 24: 232-234
  Thieme ConnectPubMedSearch in Google Scholar
• 2 Apel D, Jakobs R, Spiethoff A et al. Follow-up after endoscopic snare resection of duodenal adenomas. Endoscopy 2005; 37: 444-448
  Thieme ConnectPubMedSearch in Google Scholar
• 3 Sohn JW, Jeon SW, Cho CM et al. Endoscopic resection of duodenal neoplasms: a single-center study. Surg Endosc 2010; 24: 3195-3200
  CrossrefPubMedSearch in Google Scholar
• 4 Lépilliez V, Chemaly M, Ponchon T et al. Endoscopic resection of sporadic duodenal adenomas: an efficient technique with a substantial risk of delayed bleeding. Endoscopy 2008; 40: 806-810
  Thieme ConnectPubMedSearch in Google Scholar
• 5 Alexander S, Bourke MJ, Williams SJ et al. EMR of large, sessile, sporadic nonampullary duodenal adenomas: technical aspects and long-term outcome (with videos). Gastrointest Endosc 2009; 69: 66-73
  CrossrefPubMedSearch in Google Scholar
• 6 Oka S, Tanaka S, Nagata S et al. Clinicopathologic features and endoscopic resection of early primary nonampullary duodenal carcinoma. J Clin Gastroenterol 2003; 37: 381-386
  CrossrefPubMedSearch in Google Scholar
• 7 Inoue H, Endo M, Takeshita K et al. A new simplified technique of endoscopic esophageal mucosal resection using a cap-fitted panendoscope (EMRC). Surg Endosc 1992; 6: 264-265
  CrossrefPubMedSearch in Google Scholar
• 8 Tada M, Murakami A, Karita M et al. Endoscopic resection of early gastric cancer. Endoscopy 1993; 25: 445-450
  Thieme ConnectPubMedSearch in Google Scholar
• 9 Dixon MF. Gastrointestinal epithelial neoplasia: Vienna revisited. Gut 2002; 51: 130-131
  CrossrefPubMedSearch in Google Scholar
• 10 Shioji K, Suzuki Y, Kobayashi M et al. Prophylactic clip application does not decrease delayed bleeding after colonoscopic polypectomy. Gastrointest Endosc 2003; 57: 691-694
  CrossrefPubMedSearch in Google Scholar
• 11 Spira IA, Ghazi A, Wolff WI. Primary adenocarcinoma of the duodenum. Cancer 1977; 39: 1721-1726
  CrossrefPubMedSearch in Google Scholar
• 12 Ross RK, Hartnett NM, Bernstein L et al. Epidemiology of adenocarcinomas of the small intestine: is bile a small bowel carcinogen?. Br J Cancer 1991; 63: 143-145
  CrossrefPubMedSearch in Google Scholar
• 13 Onoe S, Katoh T, Shibata Y et al. A clinicopathological study of our 10 cases with primary carcinoma of the duodenum [in Japanese with English abstract]. Jpn J Gastroenterol Surg 2006; 39: 1458-1463
  CrossrefPubMedSearch in Google Scholar

Corresponding author

• References

• 1 Obata S, Suenaga M, Araki K et al. Use of strip biopsy in a case of early duodenal cancer. Endoscopy 1992; 24: 232-234
  Thieme ConnectPubMedSearch in Google Scholar
• 2 Apel D, Jakobs R, Spiethoff A et al. Follow-up after endoscopic snare resection of duodenal adenomas. Endoscopy 2005; 37: 444-448
  Thieme ConnectPubMedSearch in Google Scholar
• 3 Sohn JW, Jeon SW, Cho CM et al. Endoscopic resection of duodenal neoplasms: a single-center study. Surg Endosc 2010; 24: 3195-3200
  CrossrefPubMedSearch in Google Scholar
• 4 Lépilliez V, Chemaly M, Ponchon T et al. Endoscopic resection of sporadic duodenal adenomas: an efficient technique with a substantial risk of delayed bleeding. Endoscopy 2008; 40: 806-810
  Thieme ConnectPubMedSearch in Google Scholar
• 5 Alexander S, Bourke MJ, Williams SJ et al. EMR of large, sessile, sporadic nonampullary duodenal adenomas: technical aspects and long-term outcome (with videos). Gastrointest Endosc 2009; 69: 66-73
  CrossrefPubMedSearch in Google Scholar
• 6 Oka S, Tanaka S, Nagata S et al. Clinicopathologic features and endoscopic resection of early primary nonampullary duodenal carcinoma. J Clin Gastroenterol 2003; 37: 381-386
  CrossrefPubMedSearch in Google Scholar
• 7 Inoue H, Endo M, Takeshita K et al. A new simplified technique of endoscopic esophageal mucosal resection using a cap-fitted panendoscope (EMRC). Surg Endosc 1992; 6: 264-265
  CrossrefPubMedSearch in Google Scholar
• 8 Tada M, Murakami A, Karita M et al. Endoscopic resection of early gastric cancer. Endoscopy 1993; 25: 445-450
  Thieme ConnectPubMedSearch in Google Scholar
• 9 Dixon MF. Gastrointestinal epithelial neoplasia: Vienna revisited. Gut 2002; 51: 130-131
  CrossrefPubMedSearch in Google Scholar
• 10 Shioji K, Suzuki Y, Kobayashi M et al. Prophylactic clip application does not decrease delayed bleeding after colonoscopic polypectomy. Gastrointest Endosc 2003; 57: 691-694
  CrossrefPubMedSearch in Google Scholar
• 11 Spira IA, Ghazi A, Wolff WI. Primary adenocarcinoma of the duodenum. Cancer 1977; 39: 1721-1726
  CrossrefPubMedSearch in Google Scholar
• 12 Ross RK, Hartnett NM, Bernstein L et al. Epidemiology of adenocarcinomas of the small intestine: is bile a small bowel carcinogen?. Br J Cancer 1991; 63: 143-145
  CrossrefPubMedSearch in Google Scholar
• 13 Onoe S, Katoh T, Shibata Y et al. A clinicopathological study of our 10 cases with primary carcinoma of the duodenum [in Japanese with English abstract]. Jpn J Gastroenterol Surg 2006; 39: 1458-1463
  CrossrefPubMedSearch in Google Scholar