Streptozocin (STZ)-based chemotherapy has been used for over 30 years in the treatment of neuroendocrine tumours (NET); however, there have been few randomised trials in homogeneous and well-characterised patient populations. With the recent approval of sunitinib and everolimus for pancreatic NET (PNET) and the emergence of a more stratified approach to cancer therapy, it is timely to reevaluate the role of chemotherapy. Here we review the evidence base for STZ-based chemotherapy, the toxicity associated with treatment and the role of predictive markers such as Ki67 to select patients who may benefit most from therapy. Although there are no trials comparing chemotherapy with best supportive care, there is evidence that multi-agent STZ-containing regimens are associated with improved survival compared with control therapy. Compared with other therapies, chemotherapy appears to be associated with the highest response rate, particularly in PNET and remains the first-line treatment of choice for those patients in whom response is required. This includes those who are symptomatic from tumour burden and those with locally advanced disease who may be down-staged for resection. The role of Ki67 and other predictive markers requires further assessment in prospective studies as does the relative efficacy of alternative agents such as temozolomide.