This is an IRB approved, HIPAA compliant retrospective study. The medical and imaging records of 21 consecutive patients who underwent endovascular portal vein recanalization and stent placement between January 2012 and March 2020 were reviewed. Patients included 14 women and 7 men with a mean age of 66 (48–83) years. Portal vein obstruction was identified by contrast enhanced CT in all cases. Cases were presented at the multidisciplinary Tumor Board Conference to determine if patients were suitable candidates for intervention and decision to proceed was by group consensus. The clinical diagnoses were pancreatic cancer (n = 19), colon cancer metastatic to the liver (n = 1) and cholangiocarcinoma (n = 1). The etiology of venous obstruction included extrinsic compression by unresectable tumor (n = 9), portal vein stenosis after Whipple with venous reconstruction (n = 7), Yerdel class IV acute portal vein thrombosis after Whipple with venous reconstruction (n = 2), extrinsic compression and erosion of the portal vein by abdominal fluid collection after Whipple (n = 1) and Yerdel class IV acute portal vein thrombosis after partial liver resection (n = 2). Clinical presentation included: ascites (n = 5), abdominal pain, portal vein thrombosis, abnormal liver function tests and ascites (n = 4), abdominal pain and ascites (n = 4), abdominal pain and diarrhea (n = 3), gastrointestinal bleed (n = 3) an abdominal pain (n = 2). All patients signed written informed consent for the procedure.
Procedural Technique
The procedures were performed in a state of the art interventional radiology suite by two board certified interventional radiologists with more than 10 years of experience. Conscious sedation with midazolam and fentanyl was used in 18 patients and general anesthesia in 3. Broad spectrum prophylactic antibiotics with coverage for Gram negative bacteria were administered to all patients.
Access into the portal venous system was obtained using a transhepatic approach in 20 patients and a trans-splenic approach in one. Access was ultrasound guided, using a fine needle access system (Merit S-MAK, Merit Medical, South Jordan, UT). Entry into the portal venous system was confirmed by injection of non-ionic contrast through the needle. Then, a 0.018” guidewire was advanced followed by tract dilation and placement of a 5 Fr dilator. A direct portogram was then performed to confirm obstruction of the portal venous system; a 0.035” Benston guidewire (Cook, Bloomington, IN), was advanced followed by placement of a vascular access sheath (Pinnacle, Terumo, Japan). Figure # 1 illustrates portal vein access technique.
The obstruction of the portal vein was crossed using a combination of a Bernstein catheter (Angiodyamics, Latham, NY) and a stiff type glidewire (Terumo, Japan). Once the obstruction was crossed, a direct portogram using a measuring catheter (Cook Medical, Bloomington, IN) was performed in preparation for stent placement.
Four patients had Yerdel class IV acute thrombosis of the portal vein and required additional intervention including mechanical thrombectomy with Angiojet Zelante device (Boston Scientific, Natick, MS) (n = 2), thromboaspiration with the Indigo Cat-8 system (Penumbra, Alameda California) (n = 1) and power-pulse infusion of 4 mg of alteplase followed by mechanical thrombectomy with the Angiojet Zelante device (Boston Scientific, Natick, MS) (n = 1).
Stent Selection And Placement Technique
Stent choice was based on operator’s judgment and venographic findings. The stent length was selected based on the portogram with measuring catheter and the diameter was determined by measurements obtained from the CT scans. Bare self-expandable nitinol stents were used in 17 patients (Zilver, Cook, Bloomington IN and Protege EV-3, Minneapolis, MN), a balloon-expandable stent was used in 1 (Express, Boston Scientific, Natick MS), a balloon expandable stent-graft in 1 (Atrium Medical, Merrimack, NH) and a combination of self-expandable stent and a self-expandable stent-grafts (Viabahn, WL Gore, Phoenix, AZ) in 2 patients. Stents were dilated to profile using standard angioplasty balloons. Stent diameters ranged from 8–14 mm and lengths ranged from 1.7- 8 cm depending on diameter and length of the obstruction. A completion portogram was performed to confirm stent patency in all cases. If completion portogram showed residual thrombus, this was treated with additional intervention including mechanical thrombectomy, pulse spray thrombolysis and/or additional angioplasty. The transhepatic tracts were closed using 0.035” Nester coils (Cook medical, Bloomington, IN) (n = 16), and Mynx vascular closure device (Cardinal Health, Dublin, OH) (n = 4). The trans-splenic access was closed with an 8 mm Amplatzer plug (Abbott, St. Paul, MN).
Technical success was defined by successful crossing of the obstruction, stent placement and dilatation and a completion portogram showing a patent portal vein with patent intrahepatic portal vein branches. Clinical success was defined by resolution of ascitic fluid and bowel wall edema by cross sectional imaging, symptomatic improvement in abdominal pain and/or bleeding control.
Management After Intervention And Follow-up
Patients were followed from date of stent insertion until latest clinical follow-up or patient’s death (2 weeks- 4 years). All patients were seen one week after intervention in the IR clinic for recognition of early complications. An abdominal ultrasound was done within a week after the procedure to assess stent patency.
All patients were followed in the surgical oncology clinic. Contrast enhanced CT scans were performed every 3–6 months to monitor disease progression. Indications for imaging studies was decided by the medical oncologist in consultation with the hepatobiliary surgeon and the interventional radiologists.. The follow-up scans were reviewed by the interventional radiologist. If a stent abnormality was identified, the case was discussed and the possibility of stent revision was considered. Figures 2a-g illustrate the process of stent placement and follow-up in a patient who presented with acute portal vein thrombosis.
Statistical Analysis
Study results are presented in means and percentages. Statistical analysis was performed using SPSS version 26. Stent patency and patient survival are presented with Kaplan-Meier curves.