Malignant gastric outlet obstruction (MGOO) occurs as a complication in 10%–38% of pancreatic cancers.¹ Surgical gastrojejunostomy is a treatment option and is associated with a morbidity rate of 11%–22%,^2–4 whereas self-expanding metal stents (SEMS) have become the mainstay of palliation in patients with poor Eastern Cooperative Oncology Group (ECOG) performance status (PS).^4,5 The benefits of stenting are prompt symptomatic recovery and a short hospital stay. Complications include occlusion (9%–26%), bleeding (2%–6%), pain (2%–8%), biliary obstruction (2%–6%), and perforation (0.5–5%).⁶ Stent displacement is a recognised event and is influenced by the type of device (uncovered, 0%–6%; covered, 9%–37%).^6–12 However, these figures do not take into account recent advances in stent construction.¹³ This report describes the unexpected early migration of a modern partially covered duodenal stent and the role of surgery, in the context of multidisciplinary management, as a remedy and palliative measure.

A 71-year-old female presented with a 3-month history of epigastric pain, vomiting, and 10-kg weight loss, with an ECOG PS of 1. Past medical history included hypertension and breast cancer. A computed tomography (CT) scan revealed a 2.5-cm mass in the pancreatic uncinate process with superior mesenteric artery involvement and liver metastases (cT4N1M1). An endoscopic ultrasound-guided biopsy diagnosed a pancreatic ductal adenocarcinoma, and the patient was started on cisplatin and gemcitabine due to previous breast cancer history. After 4 months, she was admitted with biliary sepsis and an ECOG PS of 2, and a percutaneous transhepatic cholangiogram showed an occlusion of the distal common bile duct (CBD). A 10 mm × 60 mm double, partially covered EGIS® stent (S&G Biotech, Yongin, Korea) was placed from the distal CBD into the second part of the duodenum. Despite satisfactory stent position, inadequate drainage persisted due to a tight stricture at the third-fourth part of the duodenum (Fig. 1). An 8-Fr internal/external biliary drain (Bioteq, Taipei, Taiwan) was inserted pending duodenal stenting with a 24 mm × 100 mm partially covered double EGIS® stent, deployed just beyond the ampulla to the duodenojejunal flexure (Fig. 2). Free passage of contrast medium was seen, but initial expansion was only 25%; hence radiologic surveillance was commenced. A radiograph on day 2 demonstrated decompression of the duodenum but persistent poor expansion (25%) of the proximal end of the stent, which had also migrated distally 2 to 3 cm. On day 10, the patient complained of severe intermittent abdominal pain, and her oral intake was limited to fluids (Gastric Outlet Obstruction score [GOOS] = 1). On a CT scan, the stent had completely dislodged and migrated to the ileocaecal junction (Fig. 3). On physical examination, the abdomen was soft and non-distended, with mild tenderness in the right iliac fossa; laboratory values were normal except for the bilirubin level, which nonetheless continued a decreasing trend following biliary stenting. The treatment strategy included a combined or two-step approach. Given the patient’s stable clinical condition and her desire to minimise hospitalisation, an exploratory laparotomy was undertaken to address stent migration and duodenal obstruction. Intraoperatively, the stent was located in a mildly distended ileum, proximal to the ileocaecal junction, with no evidence of serosal compromise. A 4-cm longitudinal enterotomy was performed at 20 cm from the ileocaecal junction (Fig. 4A). The uncovered stent ends were adherent to the mucosa, whose integrity was confirmed at inspection, and the stent was safely extracted through the enterotomy (Fig. 4B). A retrocolic isoperistaltic gastrojejunostomy was fashioned after bringing a loop of jejunum through the transverse mesocolon. On postoperative day 9, the patient resumed soft oral intake (GOOS = 2), which she tolerated for 3 months until passing away.

Figure 1. Contrast-enhanced computed tomography coronal view demonstrating tumour involvement of the third part of the duodenum (white arrow), just distal to the biliary stent. The proximal duodenum (D) is dilated, and there is gas in the bile duct (black arrow).

Figure 2. (A) Contrast injection demonstrates complete obstruction of the third portion of the duodenum (arrow). The bile duct remains dilated (arrowheads). (B) Deployed stent (arrowheads). Although there is only limited expansion, the contrast has passed freely.

Figure 3. Contrast-enhanced computed tomography axial view demonstrating the duodenal stent impacted at the ileocaecal junction. There is minor dilatation of the segment containing the stent, but the proximal small bowel is of normal calibre.

Figure 4. Intraoperative findings. (A) Enterotomy (arrow) proximal to the terminal ileum containing the stent (arrowheads). (B) Retrieved duodenal stent.

This patient developed a type-III bilioduodenal stricture in the context of metastatic pancreatic cancer.¹⁴ The priority was to relieve symptoms and maximise the quality of life. Here we discuss the rationale for decision-making (Fig. 5).

Figure 5. Patient management algorithm. Bold lines indicate the pathway followed for the patient presented in this case report. BSC, best supportive care; cSEMS, covered self-expanding metal stent; ECOG, Eastern Cooperative Oncology Group; EUS, endoscopic ultrasound; GA, general anaesthesia; GJY, gastrojejunostomy; i.o., intraoperative; MDT, multidisciplinary team; MGOO, malignant gastric outlet obstruction; OTW, over-the-wire delivery system; pcSEMS, partially covered self-expanding metal stent; PS, performance status; QoL, quality of life; TPN, total parenteral nutrition; TTS, through-the-scope delivery system; uSEMS, uncovered self-expanding metal stent.

Duodenal obstruction: surgery versus stenting

Enteral stents, placed under endoscopic or fluoroscopic guidance, have been established as an effective, low-morbid alternative to gastrojejunostomy and are considered the best standard of care in the palliative setting. A 2007 meta-analysis, including one randomised trial and eight retrospective studies (n = 307 patients), reported higher clinical success, shorter time to oral intake, and less morbidity than gastrojejunostomy.¹⁵ In 2018, a Cochrane review including three randomised trials (n = 84 patients) confirmed that duodenal stenting is associated with quicker resumption of oral intake, despite higher rates of symptomatic recurrence.¹⁶ According to the SUSTENT study, gastrojejunostomy represents the best option when patient life expectancy exceeds 2 months and ECOG PS represents the most reliable prognostic factor in stratifying patients and allocating treatments (i.e., ECOG PS 0–1, gastrojejunostomy; ECOG PS 3–4, duodenal stenting).^5,17 Finally, laparoscopic gastrojejunostomy is a valuable option, although its superiority has not been confirmed over stenting.¹⁸ In our case, duodenal stenting was preferred because of the patient’s deteriorated PS, short life expectancy, and desire to minimise the hospital stay.

Stenting: insertion technique and type of device

A combined endoscopic-fluoroscopic approach has the advantages of reduced radiation dosage and procedural time, and increased stability during deployment.^7,19 The risk of obstruction and migration guides stent choice. Uncovered devices achieve firm anchorage but are affected by a higher obstruction rate; conversely, covered stents are more resistant to obstruction but prone to migration, as demonstrated by dislocation rates of 0%–6% and 9%–37%, respectively.^6,9–11,20 In the duodenal region, both uncovered and partially covered stents have fewer complications than fully covered stents and are considered the first choice in patients with MGOO.⁹ More recently, partially-covered “knitted” (D-weave, hooking wire) stents, where the structure minimises the straightening forces, allowing the device to sit coaxially in the bowel lumen, have been developed to reduce both migration and tumour ingrowth.²¹ Whilst first-generation “braided” stents had a higher axial rigidity, modern “knitted” stents may have no straightening forces at all (Fig. 6).²² Finally, fixation using an endoscopic clip has been explored in pilot studies but, due to the improved characteristics of new-generation stents, has not found application in routine practice.²³ We acknowledge that in our patient, early clip fixation, dilatation, or repositioning may have been helpful in preventing further stent migration. Nonetheless, this was an unexpected event. The patient was not on active anticancer treatment, which is a known risk factor.²⁰ Moreover, the inserted device was a partially covered double-layer knitted stent (a covering expanded polytetrafluoroethylene membrane sandwiched between two layers of nitinol skeleton) and its large (24 mm) diameter coupled with its outer metal surface and uncovered ends should have provided ample fixation. It was deployed in a good position with the proximal end beyond the ampulla and the distal end beyond the duodenojejunal flexure. As a result, the persisting poor expansion—possibly caused by the bulky tumour and tight stricture—may be individuated as the leading cause of migration. Finally, we cannot exclude more subtle reasons for failure pertaining to the intimate interaction between bowel mucosa, cancer tissue, and stent fabric. In a retrospective series including 220 patients with duodenal obstruction, the migration rate of partially covered SEMS was only 2.4%.⁹

Figure 6. Impact of different stent coverings on duodenal/colonic stents. Left, two fully silicone-covered stents of high axial rigidity. Right, two partially covered knitted stents of high conformability.

Management of migration

Despite the advent of modern stents, distal (0%–23%) or proximal (0%–17%) migration still occurs,^9,24,25 and can cause significant morbidity due to pain, obstruction and, even mortality secondary to bleeding or perforation.^8,26,27 Specialist follow-up is thus paramount, as there have also been increasing numbers of corrosive stent failures due to extended dwell times.^28,29 Migrated duodenal stents are unlikely to pass through the ileocaecal valve, with the inherent risk of bowel injury or obstruction; hence, surgical intervention is usually required, although it has to be balanced against the patient’s frailty and life expectancy.¹² Endoscopic retrieval is feasible with oesophageal stents; however, most duodenal devices migrate beyond the reach of a standard endoscope.³⁰ In our case, prompt surgical management was deemed the most effective option given the coexisting duodenal obstruction.

Management of persistent malignant gastric outlet obstruction

Although delayed re-stenting or endoscopic ultrasound-guided gastrojejunostomy could be an option, decision-making favoured a combined approach within the same procedure, in line with the patient’s preferences.

In conclusion, we report the challenging complication of a partially covered knitted duodenal stent that migrated to the ileocaecal junction. Prompt laparotomy with both stent retrieval and creation of a gastrojejunostomy proved to be a safe remedy and an effective palliative measure.

The authors would like to acknowledge Dr Mahesh Trivedi and Dr Nisha Thambuchetty for their role in patient care.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

No potential conflict of interest relevant to this article was reported.