Ductal pancreatic carcinomas are currently the fourth most common fatal cancer disease with a survival rate of less than 5 % when all stages are considered. Other malignant pancreatic tumours have markedly better prognoses. Even after complete resection and adjuvant chemotherapy, the 5-year survival rate amounts to merely 20 - 25%. Besides a high resistance to chemotherapy and early lympho- and haematogenic metastases, the reason for this is often tumour extension beyond the medial and dorsal resection margins. In standardised pathological examinations cancer cells can be detected in the resection margins in about 75 % of the cases, which reflect the aggressive and infiltrative tumour growth and probably explains the high rate of local recurrence. Standard operations for curative tumour resection are the pylorus-preserving pancreatoduodenectomy (PPPD) and the left pancreatic resection with splenectomy in cases of pancreas tail tumours. In high-volume centres the mortality can be reduced to under 3 % and the long-term survival improved with an increase of the resection rate. Considering surgical complications, pancreatic fistulas with a prevalence of up to 10 % play a decisive role. The technique of the pancreatic anastomosis as well as closure of the pancreatic tail thus represents major surgical challenges. In view of the high recurrence rate of pancreatic carcinomas, extended surgical procedures have been examined in numerous studies. Although infiltration of the portal vein is not a contraindication for curative resection with vascular reconstruction which gives comparable survival rates, a radical, extended lymphadenectomy does not seem reasonable on the basis the available data. In selected, individual cases, patients may benefit from neoadjuvant radiochemotherapy to down-stage an unresectable tumour with subsequent tumour resection, a metastasis resection, or a resection of a local recurrence. An R0 resection and tumour-free lymph nodes (N0 stage) are the two factors that can provide the best prognosis for the patient with a median survival of 2 years and a good quality of life. Pancreas surgery is being increasingly oriented to the evidence-based data from randomised, controlled studies. In order to achieve a further and urgently needed improvement in treatment results, one should consider, if possible, all suitable patients for enrolment in current clinical studies on neoadjuvant, surgical, or adjuvant therapy.