The area most infiltrated by the tumor is the margin toward the SMA [37]. Because of this, the “artery-first” procedure was developed. In general, before mobilizing the pancreatic head, the SMA is identified and dissected alongside its anatomical location. Then the neck of the pancreas and the stomach is separated. For this approach, several techniques have been proposed so far. Some start with a left-sided and others with a right-sided arterial resection and some with a supracolic approach [38–42]. Difficulties with the technique can occur in obese patients and with large tumors. In most cases, the SMA is reconstructed with a saphenous vein graft; however, transposition of the splenic artery is also an option. In addition, the “artery-first” approach has not only been proposed in conventional pancreaticoduodenectomies but also in field of minimal invasive surgery [43].

The “artery-first” technique allows identifying true involvement of the SMA before committing an irreversible step during surgery. It also provides a good balance between achieving tumor-free resection margins and preserving the nerve plexus around the SMA, which is important to avoid postoperative diarrhea. Except for the latter reasons, the approach also provides an early vascular control of the SMA and the SMV.

One of the common hepatic artery variations is the replacement of the right hepatic artery, which directly arises from the SMA (Michels III) and is located behind the pancreatic head. In addition, the common HA can also be completely replaced by a vessel, which directly originates from the SMA (Michels IX) [44–46]. These conditions should be known preoperatively. The “artery-first” approach allows early identification and handling of vascular anomalies [47].

Distal splenopancreatectomy with resection of the celiac trunk (CT) or the common hepatic artery (HA) is performed more frequently in borderline resectable PDACs. Several modifications of the original Appleby’s procedure have been published [48, 49]. The CT or HA is resected in approximately 20% of all arterial resections performed for advanced PDACs [50]. The challenge of resecting the CT or HA is to simultaneously sustain perfusion of the liver, which should be taken over by the collateral flow through the pancreaticoduodenal arcade and the gastroduodenal artery. When performing Appleby’s procedure, a constant monitoring of the vascularization of the liver is mandatory. This can be achieved by assessment of the color and tension of the liver supported by intraoperative Doppler ultrasound. The arterial flow should be greater than 22 cm/s in order to prevent liver ischemia, postoperative liver failure, and infection, which could lead to biliary complications [51, 52]. The celiac trunk can be resected down to its origin from the aorta. In case of insufficient perfusion by the collaterals or necessity of total pancreaticoduodenectomy due to oncologic reasons, reconstruction is possible as long as the proper HA can be preserved. Reconstruction can either be done by primary tension-free end-to-end anastomosis or by using a venous, arterial, or synthetic graft [48]. The right gastric and gastroepiploic arteries should be conserved to secure adequate perfusion of the stomach.

Few studies with a small numbers of patients included have evaluated the outcome after arterial resection [13, 37, 50, 53]. In general, the analyzed data is heterogeneous, and most of the conducted studies were non-randomized and retrospective. So reliable data on this topic are still missing. The meta-analysis of Mollberg et al. demonstrated that arterial resection is associated with increased perioperative morbidity and mortality as well as with a reduced 1-year survival as compared to the patients with venous vascular resection. Though, patients with advanced PDAC and arterial resection showed a better survival as compared to patients with advanced PDAC without any surgical intervention [13, 53]. Another study supports the results of the latter meta-analysis in regard to more tumor-free margins and better patient survival after arterial resection [37]. In summary, arterial resection should only be performed in high-volume centers by experienced surgeons and only in selected patients. Thereby, similar survival rates between vascular resected and nonvascular resected patients of up to a median of 18 months can be achieved [54, 52].

Resection of the superior mesenteric and portal vein has gained concurrent acceptance to achieve tumor-free margins in centers for pancreatic surgery over the last decades. If suspicion of venous involvement has been raised in preoperative imaging, the subsequent vessel should be treated as truly involved during surgery to avoid experimental dissection and potential opening of the tumor. The localization of the tumor plays an important role in further management. If tumor involvement is close to the confluence of the splenic and mesenteric vein, resection and consecutive reconstruction can be performed more easily since the lumen is large enough to create a reliable anastomosis. The farther distal the tumor is located from the confluence the more limited the technical options in resection and reconstruction of the SMV are [15].

Depending on the relation of the tumor to the vein, varying techniques in venous resections have been performed. If only minimal contact to the vessel is present, a partial venous excision with direct closure (venorraphy) or closure by a patch is suitable.

In general, if local resection of the SMV is not possible, the mesenteric root should be mobilized completely by dissecting the retroperitoneal adhesions of the right hemicolon [55]. This provides great maneuverability of the SMV and consecutively improves alignment for anastomosis.

Segmental resection with primary venovenous anastomosis or segmental resection with interposed venous graft is necessary if tangential resection of the vein is not possible [56, 57]. The clamp time of the SMV and PV should be kept to a minimum to reduce edema of the bowel. The vascular graft can either be autologous (e.g., saphenous vein) or synthetic (e.g., goretex). Implanting a synthetic graft saves operation time, since no additional venous harvesting is needed, but bears the risk of every synthetic material in regard to infection and anastomotic leakage. Especially in the setting of an increased risk for developing pancreatic fistulas, graft infection can become chronic and eventually lead to hemorrhage. Though, in a series of 110 patients, no differences in morbidity and mortality between different types of portal vein reconstruction techniques were observed [58]. Since more evidenced-based data of venous reconstructions are needed, the ISGPS proposed a classification of venous resections to standardize nomenclature for improved recommendations in the future (see Table 15.2) [15].

In two systematic reviews, no differences in morbidity and mortality have been observed for patients undergoing pancreaticoduodenectomy with and without venous resection [25, 59]. The latter studies demonstrated 3-year survival rates of 16–19.4% and 5-year survival rates of 7–12.3%, which are better than palliative treatment alone. The most important factors for survival were depth and the length of venous involvement and tumor-free resection margins. Patients with less than 3 cm of tumor involvement yielded a better survival than patients with more than 3 cm of tumor attachment. In addition, true involvement of the intima seems to be a factor of poor prognosis even in R0 resections [60–62]. After pancreaticoduodenectomy with venous resection, patients should receive adjuvant chemotherapy to further improve survival rates [63].

Venous resection during surgery for pancreatic cancer has become a standard procedure in recent years and has been included in several guidelines for pancreatic surgery. It should be performed on a routine basis by experienced surgeons if complete resection of the tumor can be achieved.