Pancreatic pseudocyst drainage in children by image-guided cystogastrostomy and stent insertion

The primary endpoint of technical success was defined as successful placement of a cystogastric stent. The secondary endpoints were complications, categorised using Society of Interventional Radiology criteria [6], and outcome, defined by residual pseudocyst size at imaging and symptoms at clinical follow-up. We used review of the medical records up to post-procedure discharge from hospital to identify and categorise delayed complications.

We analysed demographics, technical details and complications using descriptive statistics.

Children were referred for radiologic drainage of large, non-infected, symptomatic or complicated pancreatic pseudocysts. The presence of a pseudocyst was confirmed by imaging studies. Pre-procedure imaging included CT or MRI and ultrasound (US) to provide information on size, location and number of cysts (Fig. 1). As at other institutions [7], children with symptomatic pseudocysts are referred for drainage but the integrity of the pancreatic duct is not routinely imaged by endoscopic retrograde cholangiopancreatography or MR cholangiopancreatography. The decision regarding drainage approach was made at multidisciplinary team meetings, where children were assigned for either surgical or radiologic drainage. At our institutions the availability of interventional endoscopy and endoscopic ultrasound is limited.

All procedures were performed under general anaesthesia by one of four paediatric interventional radiologists (P.A.P., C.G., S.S., D.J.R.) with between 4 and 19 years experience. Prophylactic antibiotics were administered according to interventional radiologist preference. Epigastric US visualisation of the pancreatic pseudocyst and access planning was performed using either a Logiq E9 (GE Healthcare, Waukesha, WI) or Acuson S3000 (Siemens Healthcare, Munich, Germany). Percutaneous access to the stomach was performed via an existing gastrostomy tract if one existed. Otherwise, following air or saline insufflation via a nasogastric tube, the stomach was punctured under combination US and bi-plane (anteroposterior and lateral projections) fluoroscopy using an Artis Zee or Axiom Artis dBC (Siemens). To facilitate stomach insufflation, glucagon (0.03 mg/kg up to a maximum dose of 1 mg) was administered. Two or three gastropexy anchors (Cope Suture Anchor Set, Paediatric Configuration; Cook Medical, Bloomington, IN) were delivered according to radiologist preference. Combined US, fluoroscopic or cone-beam CT guidance was used to access the pancreatic pseudocyst with an 18-gauge trocar needle through the posterior wall of the stomach (Fig. 1). Iohexol (Omnipaque 240; GE Healthcare) was injected to confirm intra-pseudocyst position. A 0.035-in. stiff guidewire (Amplatz Extra-Stiff Wire Guide or Amplatz Super-Stiff Wire Guide; Cook Medical) was advanced into the pseudocyst (Fig. 1). Over the wire, the cystogastric tract was dilated using a 5-mm or 8-mm angioplasty balloon. In some cases, depending on radiologist preference, an 11-French (Fr) or 12-Fr sheath was advanced into the pseudocyst over the wire. A self-expanding covered nitinol stent designed for pancreatic pseudocyst drainage (Niti-S Nagi; Taewoong, Gyeonggi-do, South Korea) was deployed (Fig. 1). The stent size was chosen based on pre- and intra-procedural assessment of distance from gastric lumen to pseudocyst lumen. The stent was deployed under US and fluoroscopic guidance. After deployment of the distal flange, the delivery sheath was gently and slowly pulled so the distal flange was aligned with the internal cavity of the cyst. The rest of the stent was then deployed. Through the stent, the pseudocyst was aspirated. If there was a pre-existing gastrostomy, this was replaced. If there was no pre-existing gastrostomy, the gastrostomy tract formed to perform the drainage was allowed to close. If used, gastropexy sutures were cut approximately 2 weeks after the procedure.

Image-guided cystogastrostomy was performed in six children (four male), ages 46 months to 15 years (median age 6 years), with weight between 13.8 kg and 47 kg (median weight 18 kg; Table 1). The cause of the pseudocyst was chronic pancreatitis in two children (associated with hypermobility-type Ehlers–Danlos syndrome in one child), acute pancreatitis in two children (10 weeks following heart transplant in one and following commencement of asparaginase for acute lymphoblastic leukaemia in another), recurrent pancreatitis in one child secondary to a serine protease 1 (PRSS1) mutation, and blunt-traumatic transection of the pancreas in one child.

Two children had a prior failed attempt at surgical or endoscopic drainage. In the child who underwent surgery, 10 weeks prior to the image-guided pseudocyst drainage procedure there was re-accumulation. The other child had an attempted endoscopic drainage 5 weeks prior to the image-guided pseudocyst drainage procedure but the attempt was abandoned because a puncture of the pseudocyst was deemed impossible. CT (4), MR cholangiopancreatography (1) and both CT and MR cholangiopancreatography (1) performed a median of 8 days (range 2–36 days) before the procedure showed median maximum cyst diameter of 11.5 cm (range 4.7–15.5 cm).

Two children (patient 1 and patient 3) had existing gastrostomies, which were used for initial access to the stomach. A new puncture was made in the other four children following deployment of gastropexy sutures in two children. Five children received prophylactic broad-spectrum antibiotics at induction of general anaesthesia. One of the two children with a pre-existing gastrostomy and all children in whom a new stomach puncture was required received antibiotics.

A self-expanding covered nitinol pancreatic pseudocyst drainage stent (Niti-S Nagi) was successfully placed under image guidance in all children. There were no immediate or delayed complications. Children were discharged a median of 11 days (range 5–70 days) post-procedure. The longest post-procedure inpatient stay of 70 days was the child being treated for acute lymphoblastic leukaemia. Pseudocysts completely resolved in five of the six children. Follow-up cross-sectional imaging by CT (n=5) or MR cholangiopancreatography (n=1) was performed at a median of 54 days (range 12–98 days). Four children had further imaging follow-up by US (n=2) or CT (n=2). Initial imaging follow-up showed complete pseudocyst resolution in four children (Fig. 1). There were residual pseudocysts in two children measuring 2.1 cm and 3.5 cm. The 3.5-cm residual collection, seen on early CT follow-up 12 days post-procedure, had resolved before CT 216 days post-procedure (Table 1, Patient 5). The child with a 2.1-cm residual collection did not undergo further imaging follow-up (Table 1, Patient 2). There was no re-accumulation at long-term imaging performed at a median of 209 days (range 67–583 days). At last clinical follow-up, a median of 723 days (range 55–967 days) after the procedure, pseudocyst-related symptoms had resolved in all children. Two stents were removed endoscopically 55 days and 237 days after insertion, and one was noted to be absent at CT 96 days post-procedure (Patient 2) and was presumed to have passed naturally per rectum. Three stents remained in situ.