Endoscopic ultrasound guided fine needle aspiration (EUSFNA) has become the most accurate modality for characterization of pancreatic cystic and solid lesions, for differential diagnosis of indeterminate pancreatic masses and for locoregional staging of pancreatic and extrahepatic biliary tumours. EUS-FNA should also be performed in distant lymph nodes, ascites, liver, adrenal and mediastinal metastatic locations. Experienced groups reach a sensitivity over 85% with a 90-100% specificity, a positive predictive value of 98-100%, a negative predictive value of 44-80%, and an accuracy of 75-84% in evaluation of pancreatic masses. Morbidity rate (acute pancreatitis, infection, haemorrhage, perforation) is very low being around 1-2% and risk of peritoneal seeding was shown to be significantly lower than percutaneous CT guided FNA. The performance of this technique is dependent on the endoscopist and cytopathologist experience, the location, size and consistency of the tumour and the number of passes in the lesion. The type of echoendoscope or needle used does not influence the results, whereas it remains debated if presence of the cytopathologist on site might improve FNA performances. These last years, a new liquid-based cytology technique has been developed to process the specimen. Different methods exist to prepare this type of material and all these techniques improve EUS-FNA performance by decreasing the number of inadequate specimens and by increasing the possibility to obtain cell blocks allowing for ancillary techniques such as immunohistochemistry and molecular biology.