Adenocarcinoma of the pancreas is the cause of 3-4% of cancer related deaths in Italy and over 80% of all patients exhibit advanced disease. Treatment with surgery and chemio-radiotherapy may have meaningful results in resectable and locoregional tumours respectively. Chemotherapy is the treatment of choice in metastatic disease as palliative intent, although pancreatic tumour is considered resistant to treatment with conventional cytotoxicity drugs. Assessment of response of primary tumor is extremely difficult because of its anatomical location and fibrotic reaction around the tumor. Furthermore, medical problems associated with the age of patients, reduced performance status (PS), mainourished conditions, jaundice and pain, limit patients' tolerance and response to chemotherapy. 5-fluorouracil (5-FU) is the most frequently used drug in the treatment of pancreatic cancer with a RR of 28% in the trials performed in mid 1980, while more recently studies have reported a RR of 5-15%. Biochemical modulation of 5-FU by leucovorin, PALA and interferon does not appear to produce better results than 5-FU alone. 5-FU-based combination chemotherapy (FAM, SMF, etc) have shown interesting results in phase II (30-40%), but in a randomized trial the results of combination were similar to 5-FU alone (< 15%). Also, regimens containing platinum gave disappointing results just as the other combinations and cannot be recommended outside prospective clinical trials. When chemotherapy was compared to best supportive care (BSC), the results demonstrated a survival gain. Six studies, comparing chemotherapy versus BSC and 3 trials showed statistically significant difference in survival for patients treated with chemotherapy. Recently, new drugs have been introduced in the treatment of gastrointestinal tumour (gemcitabine, CPT11, raltitrexed, taxanes, etc.). Gemcitabine is a novel nucleoside analogue that has shown a very favourable toxicity profile and RR of 10-15% in advanced pancreatic cancer. Data from a phase II and randomized comparative trials suggest that gemcitabine offers an advantage over 5-FU in terms of improvement of PS and general clinical symptoms. Given the difficulty of accurate tumor measurement in this disease, some authors introduced a novel new end-point to evaluate the response: clinical benefit (CB). In a randomized trial of gemcitabine vs. 5-FU, RR using CB was 23.8 with gemcitabine and 4.8 with 5-FU, this difference was statistically significant with a median survival of 5.6 and 4.4 months, respectively. In conclusion, future studies should focus on phase III trials with gemcitabine, alone or in combination and phase II with new promising drugs. Quality of life, pharmaco-economic studies, CB should be the principal end-point of these studies. All patients with advanced pancreatic cancer should be included in clinical cooperative trials.