Vascular resection in pNEN surgery

Porto-mesenteric venous resection

Comparable to pancreatic adenocarcinomas, pNEN may locally affect vascular structures including the superior mesenteric (SMV) and portal vein (PV) or the superior mesenteric artery (SMA) and branches of the celiac axis (CA), respectively. There is no clear definition such as “borderline resectable” or “locally advanced” for pNEN, however, a venous infiltration is found in up to 33% of patients [1-3].

A special pNEN-associated feature in case of venous infiltration is the formation of – partially extended – thrombus formation, which does not automatically preclude patients from resection but imposes specific technical challenges during surgery [2]. Retrospective series on SMV/PV resection in pNEN report postoperative morbidity rates of 21-48% and mortality of <5% which reflects that these operations basically do not differ from comparable surgical procedures in pancreatic cancer patients [2-7]. Based on these factors, current ENETS guidelines recommend PV/SMV resection and reconstruction. Accordingly, a NANETS consensus paper, published in 2020, states that suspected venous infiltration in preoperative cross-sectional imaging does not imply a contraindication for surgical resection. Especially as imaging does not always reliably predict true vascular invasion, this can only be evaluated intraoperatively which underlines the importance for surgical exploration aiming at a radical tumor resection [8, 9]. With regards to neoadjuvant therapy in order to increase resectability of pNEN, no evidence-based standards are available today. PRRT has been shown to have a potential beneficial effect in terms of an improved progression-free survival if a radical R0 resection can be achieved after pretreatment [10]. Clinical practice evaluated by a recent survey among 155 surgeons of the ESSO-EYASAC and E-AHPBA community shows that mainly somatostatin analogues (81%) and cytotoxic chemotherapy (55%) are administered followed by molecularly targeted agents (29%) and PRRT (24%) [11]. Survival data for pNEN patients undergoing PV/SMV resections show median survival times of 60 months or 5-year overall survival rates of 45-66%, respectively [1, 3, 4]. A study from the US even reported overall survival of 60% after 10 years (12). These results underline the feasibility and oncological justification of such extended surgical procedures for pNEN patients. A conclusion on outcomes of various types of venous resection (ISGPS 1-4) or – if necessary – a preferable graft type (i.e., autologous or alloplastic) for reconstruction cannot be made from the available data.  

Venous bypass first

The before-mentioned existence of a tumor-associated PV thrombosis or a complete PV occlusion by the tumor itself can lead to portal hypertension and cavernous transformation (Fig. 1) [13, 14]. This should not be regarded as a general contraindication for surgery but may require changes in the surgical strategy. In case of pronounced venous collaterals around the pancreatic head, standard resection can lead to massive bleeding and the need for prolonged clamping of the PV / SMV or SMV branches with a high risk for small bowel congestion and consecutive ischemia. In such situations a “venous bypass first” approach is advisable [15]. Before attempting resection, the SMV or a suitable branch should be identified and a long bypass (i.e., using artificial allografts) can be created which allows a decompression of the small bowel and a resection without time pressure and excessive blood loss. Depending on the primary accessibility of the PV towards the hilum of the liver, this bypass can be interposed either between SMV and PV or between SMV and V. cava (Fig. 2) as a temporary shunt. As the V. cava is easily accessed after mobilization of the mesentery, this may be the easier approach in most patients in the beginning of the operation. After bypass creation, the resection can be carried out i.e., as an “uncinated first” approach which implies that the hepatoduodenal ligament – which often also shows venous collaterals – is dissected in a later stage of the procedure when pancreas transection and soft dissection from the SMA and CA has already been completed. As the final step, PV clamping towards the liver hilum and – if require – thrombectomy is performed (Fig. 3). Following retrieval of the specimen, the temporary SMV / V. cava shunt is reversed and SMV / PV anastomosis performed either by direct end-to-end anastomosis with complete removal of the shunt material or by using the initial prosthesis which is then shortened under preservation of the SMV anastomosis and only connected end-to-end to the PV lumen (Fig. 4). This procedure can facilitate complex resections and shows feasible outcomes in terms of morbidity and mortality without obvious differences in terms of liver enzyme elevation or failure between both shunt possibilities (SMV / V. cava vs. SMV / PV) [16].

Arterial Resections

Arterial resections for pNEN have been reported only anecdotally [3, 12]. Therefore, neither any conclusions on the feasibility, potential types of resections and arterial reconstructions nor on oncological outcome of such operations are possible to date.