Nonampullary duodenal tumors (NADTs) are reported in 0.3–4.6% of patients attending for upper gastrointestinal endoscopy [
Endoscopic management of NADTs provides a challenge in terms of accurate diagnosis, staging, and endoscopic resection in the presence of the thin duodenal wall and rich vascularity. However, endoscopic approach offers considerable advantages in terms of organ preservation, procedure-related risks, recovery, and length of hospital stay. There was a report that surgical or endoscopic resection of early duodenal cancer resulted in no lymph node metastasis in any of the cases among 128 lesions of intramucosal carcinoma [
Although there were some reports on the efficacy of endoscopic resection for NADTs from various single centers, multicenter studies have not been reported. The aim of this study was to evaluate the feasibility of endoscopic resection for the management of sessile NADTs (SNADTs) on multicenter basis.
Medical records on endoscopic resection for SNADTs were reviewed in 5 teaching hospitals affiliated to The Catholic University of Korea (Incheon St. Mary’s Hospital, Bucheon St. Mary’s Hospital, Yeouido St. Mary’s Hospital, Seoul St. Mary’s Hospital, and St. Paul’s Hospital) from July 2002 to July 2013. At least 50 cases of EMR and/or ESD for neoplasia of upper gastrointestinal tract per year are performed in every center. Patients with ampullary or periampullary lesions as well as patients with a history of familial polyposis syndromes were excluded. Pedunculated polypoid lesions were also excluded since these lesions can be easily removed by endoscopy. After reviewing the final pathologic reports acquired from endoscopic resection, adenoma, adenocarcinoma limited to the mucosal layer, and carcinoid tumors limited to the mucosa were included in this study. Demographic characteristics including sex and age and characteristics of the sessile lesions such as number, size, location, histologic findings, and endoscopic resection method were identified. The Institutional Review Board of The Catholic University of Korea approved this study.
The techniques of endoscopic resection were classified into three groups: endoscopic polypectomy (EP), which was performed by snare only without injection into submucosal layer; endoscopic mucosal resection (EMR), which was performed by snare after injection into submucosal layer; endoscopic submucosal dissection (ESD), which included the steps of precutting of mucosa and dissection of the submucosal layer with knives after injection into submucosal layer.
Bleeding was defined as intraprocedural massive bleeding that required blood transfusions or postprocedure bleeding that required blood transfusion, endoscopic intervention, or surgical intervention.
Perforation was defined when intra-abdominal space was directly observed during the procedure (frank perforation) or free air was found on a plain chest X-ray after procedure without a visible duodenal wall defect during procedure (microperforation).
Local recurrence was defined as identifying a microscopic adenoma and/or carcinoid tumor at the original tumor site during the follow-up period. Follow-up period was defined as the interval between the date of resection and the most recent endoscopic examination.
Differences in overall outcomes among the endoscopic resection methods were evaluated using the Kruskal-Wallis test or Mann-Whitney
One hundred eleven lesions from 108 patients were screened and 56 lesions from 54 patients were identified (Figure
Demographic features.
Number of patients | 54 |
Number of lesions | 56 |
Mean age (years ± SD*) | 59.5 ± 12.5 |
Male : female (%) | 33 : 21 (61.1 : 38.9) |
Histologic types (%) | |
Adenoma, low grade dysplasia | 25 (44.6) |
Adenoma, high grade dysplasia | 11 (19.6) |
Adenocarcinoma | 2 (3.6) |
Carcinoid tumor | 18 (32.1) |
Location (%) | |
1st portion | 24 (42.9) |
2nd portion | 31 (55.4) |
3rd portion | 1 (1.8) |
Size of the lesions (median (range), cm)† | |
EP | 1.2 (1.0~1.5) |
EMR | 0.8 (0.3~4.5) |
ESD | 0.8 (0.4~3.5) |
†Size of long axis.
EP: endoscopic polypectomy; EMR: endoscopic mucosal resection; ESD: endoscopic submucosal dissection.
Most of the lesions were located in 1st and 2nd portion of duodenum in a single lesion. Mean length of the long axis was 1.06 cm. On histologic examination, adenoma with low grade dysplasia was found in 25 lesions (44.7%), adenoma with high grade dysplasia in 11 lesions (19.7%), adenocarcinoma in 2 lesions (3.6%), and carcinoid tumors in 18 lesions (32.1%).
Outcomes of endoscopic resection are summarized in Table
Outcomes and complications of endoscopic resection.
EP ( |
EMR ( |
ESD ( |
|
|
---|---|---|---|---|
|
5 (100) | 35 (77.8) | 4 (66.7) | 0.414 |
Histologic complete resection (%) | 5 (100) | 37 (82.2) | 5 (83.3) | 1.000 |
Procedure time (median (range), min) | 5.0 (5~16) | 13.0 (10~130) | 41.5 (32~180) | 0.003 |
Complications (%) | 0 (0) | 3 (6.7) | 2 (33.3) | 0.140 |
Bleeding (%) | 0 (0) | 1 (2.2) | 0 (0) | 1.000 |
Perforation (%) | 0 (0) | 2 (4.5) | 2 (33.3) | 0.088 |
Follow-up period (median (range), mon) | 33.5 (7~60) | 6.0 (3~66) | 18.0 (2~96) | 1.000 |
Recurrence rate (%)* | 0/2 (0) | 1/29 (3.4) | 0/5 (0) | 0.632 |
Number of cases and complications in each center† | ||||
Incheon St. Mary’s Hospital | 0 | 4 (0) | 2 (1) | |
Bucheon St. Mary’s Hospital | 2 (0) | 10 (2) | 1 (0) | |
Yeouido St. Mary’s Hospital | 0 | 15 (0) | 3 (1) | |
Seoul St. Mary’s Hospital | 3 (0) | 15 (1) | 0 | |
St. Paul’s Hospital | 0 | 1 (0) | 0 |
EP: endoscopic polypectomy; EMR: endoscopic mucosal resection; ESD: endoscopic submucosal dissection.
*Recurrence rate was obtained from patients who were followed up at least 2 months.
†Numbers indicate the total number of procedures and the numbers in parentheses indicate the number of complications.
Immediate complication was described in 5 patients (1 bleeding and 4 perforations; Table
Median follow-up period was 33.5 months for EP, 6.0 months for EMR, and 18.0 months for ESD. Among these patients, local recurrence occurred in 1 patient who was treated with EMR. In this patient, resection margin was positive and the recurrent lesion was found 2 months after resection. This recurrence was ablated with argon plasma laser coagulation and there was no recurrence after the ablation therapy for the following 18 months.
In duodenum, primary epithelial neoplasia and carcinoid tumors are very rare, although their incidence has been increased in Korea for the past decade in part because of increased screening endoscopy and because of better awareness. Surgical and/or endoscopic resection is recommended for these lesions due to malignant potential of both lesions [
Pedunculated lesions can be easily removed by EP technique and complications such as perforation might be extremely rare. So we analyzed sessile lesions only (Paris Classification Is and II lesions) [
In this study, EP seemed to be superior compared to other techniques in the point of complete resection and complications. However, EP can only be performed in lesions with a good view and when the size is small enough to be removed without submucosal injection. ESD required statistically significant more procedure time compared to EP and EMR. However, it is plausible that ESD might be selected as a procedure of choice when
Although EMR was the most commonly performed procedure and other techniques such as EP and ESD were limitedly performed in this study, overall outcomes of endoscopic resection for SNADTs were quite favorable.
There are some limitations in this study. Although our study was a multicenter study, its retrospective design may have resulted in underreporting of complications and selection bias might have occurred inevitably. We tried to compare each endoscopic technique but patients who underwent EP and ESD were relatively small in number compared to EMR. Follow-up endoscopy was arbitrarily performed and follow-up was not evenly performed in some patients.
In conclusion, endoscopic resection for SNADT seems to be effective and safe. Additional studies including large number of cases and prospective design with long-term follow-up is anticipated.
The authors declare that there is no conflict of interests regarding the publication of this paper.
Sung Min Park and Joo Ho Ham contributed equally to this paper.