INTRODUCTION
Neuroendocrine tumors (NETs) are rare tumors that are often asymptomatic and most commonly found in the gastrointestinal tract [1-3]. Gastrointestinal NETs account for a small fraction of all gastrointestinal cancers, with the small intestine being the most common site, followed by the rectum, appendix, large intestine, stomach, and duodenum. Primary duodenal NETs (DNETs) are typically located in the first and second portions of the duodenum, with over 90% of cases occurring in these regions [4]. DNETs are increasingly recognized clinically owing to the growing use of upper gastrointestinal endoscopy and advancements in endoscopic techniques [5].
Based on their location, DNETs are classified into two main categories: periampullary and non-ampullary DNETs (NADNETs). Approximately 20% of DNETs occur near the ampulla of Vater, a region known as the periampullary area [6]. Periampullary DNETs may lead to complications such as biliary or pancreatic duct obstruction and hemorrhage and has a higher risk of spread and invasion. Therefore, patients with periampullary DNETs are commonly treated with surgical resection. The European Neuroendocrine Tumor Society (ENETS) consensus guidelines suggest that the treatment of NADNETs can be divided into three categories based on lesion size. For lesions ≤10 mm, endoscopic resection (ER) is recommended, whereas surgical resection is considered optimal for lesions >20 mm. However, the treatment strategy for lesions measuring 10–20 mm remains controversial [7,8]. According to the National Comprehensive Cancer Network (NCCN) guidelines, ER is recommended when it is locoregional and endoscopy is feasible. Duodenal local excision or pancreatoduodenectomy can also be considered [9].
NADNETs are primarily found in the first or second portion of the duodenum without the involvement of the ampulla of Vater. They are mostly nonfunctional, with approximately 90% detected incidentally during routine endoscopy [10]. For treatment, ER is generally the preferred treatment option, whereas surgery is considered only when ER is not feasible [7]. The management of NADNETs depends on several factors, such as tumor size, location, histological grade, stage, and type. However, owing to their low incidence and uncertain natural history, standardized treatment protocols have not yet been established, and there is no definite consensus [11]. This study aimed to evaluate the short- and long-term therapeutic outcomes of endoscopic and surgical treatments in patients with NADNETs.
METHODS
Study population and information
This retrospective study was conducted at Seoul National University Hospital. The medical records of patients diagnosed with NADNETs between January 2009 and December 2022 were reviewed. A total of 62 patients were diagnosed with NADNETs via endoscopic biopsies, and 7 patients were excluded due to inadequate pathological results or incomplete records (Fig. 1).
Of the 55 patients with biopsy-confirmed NADNETs, 21 underwent ER, 28 underwent surgical resection, and 6 had no residual tumor detected during follow-up after their initial biopsy. One patient had failed ER because of a non-lifting sign following normal saline injection into the submucosal layer around the lesion; consequently, surgical resection was performed. This patient was included in the surgical resection group.
All patients provided written informed consent before undergoing endoscopic or surgical treatment. The study protocol was approved by the Institutional Review Board of Seoul National University Hospital (H-2305-148-1434). The requirement for informed consent was waived by the institutional review board owing to the retrospective nature of this study.
Endoscopic and surgical treatments
Twenty-one patients underwent ER performed by experienced endoscopists at Seoul National University Hospital. Conventional endoscopic mucosal resection (EMR) with a snare and EMR with circumferential mucosal precutting (EMR-P) were performed using a conventional single-channel endoscope (GIF-H260, GIF-H290; Olympus) under conscious sedation via midazolam injection [8]. The endoscopic treatment method was selected based on the decision of the endoscopists.
Among the 21 patients who underwent ER, 20 (95.2%) were treated with conventional EMR (Fig. 2), while 1 (4.7%) was treated with EMR-P (Fig. 3). Prophylactic clipping was performed in eight cases due to deep margins at the EMR site. Of the 21 EMR cases, 2 were combined with argon plasma coagulation (APC) due to difficulties in grasping and poor lifting, which led to piecemeal resection. Preventive APC ablation was cautiously performed along the resected margins.
A total of 28 patients underwent surgical treatment, including 15 who underwent wedge resection, 1 who underwent segmental resection, 3 who underwent subtotal gastrectomy, 6 who underwent Billroth II distal gastrectomy, 1 who underwent transduodenal ampullectomy, and 2 who underwent pylorus-preserving pancreaticoduodenectomy. The surgical method was determined using a multidisciplinary approach, considering tumor size, location, extent of invasion, and comorbidities, after discussion among endoscopists, surgeons, and pathologists.
One patient who underwent transduodenal ampullectomy was initially considered to have a benign lesion but was later histologically confirmed to have a grade 1 NET. The decision to perform lymph node dissection during surgery was based on tumor characteristics, such as size, depth of invasion, differentiation, lymphovascular invasion, and the NCCN guidelines.
A preoperative evaluation was performed using endoscopic ultrasound (EUS) in 13 patients and positron emission tomography-computed tomography (PET-CT) in 9 patients. Postoperatively, most patients were followed up with CT. PET-CT follow-up was carried out in three patients based on clinical judgment, while EUS was not performed.
Histopathological examination and definition of complete resection
En bloc resection is a procedure in which the entire tumor is excised in one piece. Complete ER indicates the total removal of the tumor with no visible remnants at the resection site, regardless of whether en bloc resection was performed. Histopathological curative resection indicates no lateral or vertical margin involvement and absence of lymphovascular invasion, and it does not necessarily require en bloc resection. Histological grades were defined based on mitotic and Ki-67 indices according to the 2010 WHO classification [7].
Statistical analysis
Data were analyzed using IBM SPSS for Windows (version 25.0; IBM Corp.). Continuous data were evaluated using the Mann–Whitney U or Kruskal–Wallis tests. Categorical variables were examined using the χ2 or Fisher exact tests. A p-value <0.05 was considered statistically significant.
RESULTS
Demographic and clinical characteristics of patients
Patients diagnosed with NADNETs were included in the analysis, and their characteristics are shown in Table 1. The mean ages at the time of diagnosis were 55, 56, and 45 years in the ER, surgical resection, and biopsy-removed groups, respectively. In the ER group, 52.4% were male patients, and 47.6% were female patients. Similarly, in the surgical resection group, male patients accounted for 60.7%, whereas female patients made up 39.3%. Symptoms, such as epigastric pain, epigastric soreness, dyspepsia, and abdominal discomfort, as well as carcinoid-related symptoms were not observed in the patients.
NADNETs lesion characteristics
Baseline lesion characteristics are presented in Table 2. Most of the lesions were located in the duodenal bulb, accounting for 81.0%, 78.6%, and 66.7% in the ER, surgical resection, and biopsy-removed groups, respectively. The second portion of the duodenum followed, accounting for 19.0%, 17.9%, and 33.3% in the ER, surgical resection, and biopsy-removed groups, respectively (p=0.492). Most endoscopically observed lesions were solitary, accounting for 95.2%, 92.9%, and 100% in the ER, surgical resection, and biopsy-removed groups, respectively. The remaining cases involved two lesions.
The mean endoscopically measured tumor sizes were 9.2, 13.9, and 4.8 mm in the ER, surgical resection, and biopsy-removed groups, respectively. In the biopsy-removed group, the endoscopically measured tumor size refers to the initial size before the biopsy. During follow-up endoscopy, no visible lesions were observed, and a re-biopsy of the same area yielded a negative result. These cases included patients who refused surgical or endoscopic treatment and opted for follow-up.
When classifying the pathological tumor size into group 1 (1–5 mm), group 2 (6–10 mm), and group 3 (≥11 mm), 47.6% and 9.5% of the patients in the ER group were in group 1 and group 3, respectively. In contrast, 50.0% of patients in the surgical resection group were in group 3.
Histopathological examination associated with invasion depth and lymph node metastasis
The histopathological examinations of NADNETs are detailed in Table 3. After resection, pathological evaluation indicated that both the ER and surgical resection groups achieved curative resection with negative margins in 85.7% of cases. According to the 2010 WHO classification, tumors with a mitotic index <2 per 10 high-power field (HPF) were predominant, accounting for 76.2% and 75.0% in the ER and surgical resection groups, respectively. The mean Ki-67 indices were 1.8% and 2.4% in the ER and surgical resection groups, respectively.
A histological examination revealed that 3 tumors (1 tumor in each group) were confined to the mucosa, whereas 33 tumors extended into the submucosa: 18 (85.7%) in the ER group, 15 (53.5%) in the surgical resection group, and 0 (0%) in the biopsy-removed group. In the ER and biopsy-removed groups, no invasion beyond the muscularis propria was observed. However, in the surgical resection group, 10 (35.7%) tumors had invaded beyond the muscularis propria.
In the surgical resection group, postoperative pathological examination revealed regional lymph node metastasis in two patients. In one patient, preoperative endoscopy revealed a lesion measuring 12 mm, and its unfavorable shape prompted surgical treatment. CT imaging showed no significant lymph node enlargement, and gallium-68 DOTA-Tyr3-Octreotide (68Ga DOTA-TOC) PET revealed no definite abnormal lesions in the stomach. Postoperative pathology showed a lesion measuring 17×15 mm, with invasion into the muscularis propria, a grade 2 classification, mitotic count of 3/10 HPF, and Ki-67 index of 5%. The surgical margin was positive at the distal margin, and among the 50 examined lymph nodes, 1 metastatic lymph node was identified. According to the American Joint Committee on Cancer 8th edition pathological staging, the patient was staged as pT2N1.
In another patient, endoscopy revealed a 20-mm lesion, and preoperative 68Ga DOTA-TOC PET identified two nodular lesions in the periduodenal area with intense uptake, suggesting metastasis. Although periduodenal metastasis was possible, the lesion was assessed as localized, and surgery was performed. Postoperative pathological examination showed a lesion measuring 17×14 mm, with invasion into the muscularis propria, a grade 1 classification, mitotic count of 1/10 HPF, and Ki-67 index of 3.6%. The surgical margin was free from the tumor, but metastasis was observed in 2 out of 27 examined lymph nodes, leading to a staging of pT2N1.
Short- and long-term outcomes of ER and surgical resection
Three out of 21 (14.3%) patients in the ER group experienced perforations. One patient, with a lesion located in the duodenal bulb, was managed with hemoclips immediately after the procedure because of a high risk of perforation during the intervention. The other two patients, with lesions located in the duodenal bulb and the second portion, were treated with primary repair surgery.
Delayed adverse events, such as bleeding, stenosis, and delayed perforation, did not occur in the ER group. In the surgical resection group, there were no cases of postoperative bleeding or perforation; however, four (14.2%) patients experienced ileus, requiring an extended hospital stay. The mean hospital stay was three times longer in the surgical resection group (9 days) than in the ER group (3 days). The 30-day mortality rate, which indicates the rate of death within 30 days of a specific treatment, was 0% in both groups.
The median follow-up periods were 42 (range 12–96), 48 (range 5–85), and 42.5 (range 8–105) months in the ER, surgical resection, and biopsy-removed groups, respectively (Table 4). During the follow-up period, no local tumor recurrence or metastasis was observed in any patient.
Details of margin-positive cases in the endoscopic and surgical resection group
In the ER group, tumors were removed via en bloc ER in 19 patients, and complete ER was achieved in all patients. Information on margin-positive cases after resection is shown in Table 5. Among the three (14.3%) patients with margin-positive findings in the ER group, two had positive vertical margins, and one had both positive horizontal and vertical margins. All three patients were recommended for surgery, but they did not undergo it due to advanced age, comorbidities, or refusal of surgery. Instead, they underwent endoscopy every 6 months for 2 years and were then followed up with endoscopic examinations at 1-year intervals for a median duration of 50 months (range 4–75 months). Moreover, abdominal CT was performed annually. Surveillance biopsies were performed during endoscopy, and all revealed no remnant tumor, only the resection scar.
In the surgical resection group, four (14.3%) patients had margin-positive findings in the resection specimens; one patient had positive horizontal margins, two had positive vertical margins, and one had both positive horizontal and vertical margins. Without any additional surgery, all patients were monitored through follow-up CT and endoscopy at 1-year intervals, with no evidence of recurrence for a median duration of 46.5 months (range 33–63 months). One patient with positive horizontal margins was followed up for 33 months, two patients with positive vertical margins were followed up for 59 and 63 months, and one patient with both horizontal and vertical positive margins was followed up for 36 months.
DISCUSSION
This study analyzed the short- and long-term therapeutic outcomes of 55 patients with NADNETs who underwent ER, surgical resection, or observation after biopsy. No local recurrences or metastases were observed in any patient during the median follow-up period (range 5–105 months), even in patients with positive surgical margins. This suggests that for patients with NADNETs with lesions measuring approximately 10 mm, endoscopic treatment may offer favorable long-term outcomes comparable to those of surgical resection, which is invasive and associated with longer hospital stays. However, perforation may occur during ER regardless of tumor size, and endoscopists should exercise caution when performing ER for NADNETs.
According to treatment guidelines, the NCCN guidelines recommend ER for NADNETs measuring <10 mm. However, for lesions measuring 10–20 mm, there is no standardized treatment choice between ER and surgical resection [9]. Considering the low risk of lymph node metastasis in 10–20 mm lesions and non-aggressive course of NADNETs, minimally invasive endoscopic treatment remains the optimal first-line option.
In this study, one patient with a 13-mm grade 2 lesion (mitotic index, 1/6 HPF; Ki-67 index, 3.91%) remained stable without any recurrence for up to 16 months after ER and was subsequently monitored through regular endoscopies at an outside hospital for an additional 5 years. Similarly, three patients with positive margins after resection also showed a stable course without recurrence. In cases of grade 1 or grade 2 lesions with a mitotic index close to grade 1, surveillance can be considered without additional surgery, even if the margin is positive. The absence of recurrence in the four patients with positive margins who underwent surgical resection reflects outcomes similar to those observed in patients who underwent endoscopic treatment with positive resection margins. Our study supports the indication of ER as a viable treatment option for NADNETs, as recommended in the guidelines.
The potential for lymph node metastasis is a critical factor to consider when opting for ER as the treatment approach for DNETs [12]. Several studies have reported risk factors associated with lymph node metastasis in DNETs [13-15]; these risk factors include tumor size >10 mm, lymphovascular invasion, grade 2, and invasion into layers deeper than the submucosa. For lesions measuring 10–20 mm, the possibility of lymphovascular invasion increases, making it controversial whether endoscopic or surgical resection is the most suitable treatment approach. Determining the optimal treatment modality for 10–20 mm NADNETs is challenging; however, they can often be removed with minimal incisions using endoscopic techniques, which are less burdensome for patients than highly invasive surgery, especially for older adult [16,17]. Therefore, a multidisciplinary approach is necessary to determine the appropriate therapeutic modalities for these lesions.
According to the ENETS guidelines, additional surgery is recommended for patients with resection margin involvement [7]. However, owing to the rarity of DNETs, there is insufficient evidence to support this approach. Therefore, in clinical practice, surveillance through endoscopy and CT scans at 6–12-month intervals is preferred over invasive surgery and is widely practiced. However, the duration of follow-up after resection has not yet been standardized.
In this study, three patients required surgical resection after ER due to positive margins, but all declined surgical treatment. Instead, they underwent endoscopic and imaging follow-ups and remained stable without recurrence during a follow-up period of 50–75 months. Although there are no clear guidelines for the follow-up interval, at our institution, we performed endoscopic follow-ups every 6 months for 2 years after ER.
The reasons recurrence does not occur even when the margins are positive after ER of NADNETs can be attributed to several factors. The first is the biological characteristics of the tumors. NADNETs are often well-differentiated, with slow growth and low aggressiveness. Consequently, even if the margins are positive, the remaining cells may not actively proliferate or lead to a clinically significant recurrence [18,19]. The second factor is that the lesions were endoscopically resected completely; thus, even if there was a positive margin, it was likely to be very small. In other words, cancer cells may be detected at part of the resected tumor margin; however, because the margin thickness or width is typically minute, the risk of recurrence is unlikely to be significant. The third factor is the use of the ER technique. ER typically uses electrocautery tools that may cause thermal damage to the residual tissue, potentially reducing the risk of recurrence. Even if the resected specimen shows positive margins, residual cancer cells in the resected area may be destroyed by the heat from electrocautery [20,21].
Although en bloc resection is not feasible, piecemeal resection can be performed [22,23]. In this study, two patients had their tumors removed piecemeal with EMR, followed by ablation using APC, and both showed good long-term results during follow-up periods of 75 and 96 months. Therefore, electrocautery-based techniques such as APC can be considered an additional beneficial option when complete resection may not be achieved. However, because the duodenal wall is thin, the risk of perforation is higher than that in other areas; therefore, only experienced endoscopists should approach these procedures.
In this study, among the patients positive surgical margins, three had tumors in the second portion and four in the duodenal bulb. Perforation occurred in two patients in the duodenal bulb and one in the second portion. Therefore, it is challenging to definitively conclude that the second portion is more strongly associated with adverse events and incomplete resection based on these findings. However, one study suggested that tumor location in the second portion was associated with risk factors for non-curative resection [24].
In a previous multicenter study in Korea, the pathological complete resection rate, including EMR and endoscopic submucosal dissection, was 50% (p=0.017). However, in this study, the rate was higher at 85.7% (p=0.654). This difference is likely due to the smaller sample size, relatively smaller lesion sizes, and differences in treatment methods [12].
To reduce the risk of perforation, new techniques such as cold snare polypectomy and underwater EMR have recently been introduced and are increasingly being used [25]. Cold snare polypectomy does not require electrocautery, which reduces the risk of complications, such as perforation or delayed bleeding [26]. However, the resection depth is shallower than that in conventional polypectomy, making submucosal resection less effective. Therefore, its use in tumors with submucosal invasion increases the risk of incomplete resection.
In underwater EMR, the lumen is filled with water instead of air, which helps the mucosal layer float above the muscle layer. This facilitates resection and lowers the risk of perforation, particularly in thin-walled organs, such as the duodenum [27]. However, underwater EMR has a lower complete resection rate and may pose additional risks, such as water intoxication, requiring careful monitoring during the procedure [28].
In the current study, three patients in the ER group developed perforations and were treated without delayed adverse events. The treatment options for duodenal perforation include surgical repair, endoscopic intervention, or conservative management, depending on the patient’s condition. Given the considerable risk of perforation during ER, endoscopists should proceed with caution.
This study has some limitations. It is a single-center retrospective study, which may have introduced selection bias. Additionally, this study included a small number of patients and a limited number of grade 2 and large lesions (>20 mm). However, unlike other studies, we were able to analyze outcomes over a relatively long follow-up period. Due to the rarity of DNETs, this study is expected to contribute to the establishment of definitive NADNET treatment guidelines.
In conclusion, the recurrence-free survival of patients undergoing ER for NADNETs measuring approximately 10 mm was comparable to that of patients undergoing surgical resection. However, ER carried a relatively high risk of perforation due to difficult access and the thin duodenal wall.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
