Radical resection is currently regarded as the standard operation for pancreatic cancer. Unfortunately, in 10% to 20% of patients, extension beyond the pancreas to involve contiguous major vascular structures limits respectability and possible cure. Extending the scope of the operation to include segments of the PV and/or SMV has proved to be safe, and immediate reconstruction is feasible using well-established vascular surgical techniques. In some instances, resection of the SV–mesenteric vein confluence including the IMV is necessary to achieve an R0 resection. If the SV-IMV confluence remains intact, venous outflow from the spleen is preserved, thereby avoiding the potential development of left-sided portal hypertension and hypertensive gastropathy and/or gastric variceal hemorrhage. Ferreira et al1 have nicely demonstrated that if preservation of the SV-IMV confluence is not possible, direct anastomosis between the remaining SV and IMV segments is both feasible and safe. Furthermore, they have demonstrated short-term anastomotic patency using clinical observation of gastric venous congestion as well as color Doppler ultrasonography. The main limitations of the study are the short observation and the use of surrogate markers (platelet count and spleen volume) to assess gastric venous hypertension. While the validity of these surrogate markers of portal hypertension has been documented in patients with chemotherapy-induced hepatic sinusoidal injury and generalized portal hypertension, it is not clear whether they are equally applicable or reliable in the setting of pure left-sided portal hypertension. Other areas of concern involve the indications for and the success of such extended venous resections. Two-thirds of their patients who underwent venous resection had pathological venous involvement, and tumor involved the deeper aspects (tunica media and/or intima) of the vein wall in 60% of these patients. It has been shown elsewhere that deep invasion of the vein wall carries the same dismal prognosis as noncurative resection. Moreover, despite the extended scope of resection, an R0 resection was not achieved in 35% of patients. Ferreira and colleagues have clearly demonstrated the feasibility and safety of direct SV-IMV anastomosis when it is necessary to achieve an R0 resection for pancreatic carcinoma. However, it is our opinion that the indications for and long-term outcomes resulting from such extended venous resections will require further study and elucidation.