Carcinogenesis is a multistep process of long-term accumulation of genetic and epigenetic aberrations at the molecular level. Understanding the mechanisms underlying carcinogenesis could further provide insights for rationally designed new therapeutic strategies for cancer prevention and treatment. Epidemiological and experimental evidence supports the preventative effect of non-steroidal anti-inflammatory drugs (NSAIDs) regarding cancer development. NSAIDs inhibit cyclooxygenase (COX) activity, thus blocking the endogenous prostaglandin production. COX-2 is a key isoenzyme in this biochemical cascade and is inducible by various oncogenic stimuli. A large volume of research data has shown that COX-2 is often upregulated in many malignant tumours, rendering it an attractive candidate target for cancer therapeutics. Various possible direct mechanisms for COX-2 implication in carcinogenesis have been suggested, whilst intense interest has recently been focused on COX-2-independent effects of NSAIDs. Several COX-2-selective inhibitors are currently under evaluation in preclinical and clinical studies, either as single agents or in combination with conventional chemotherapy, radiotherapy and other new molecularly-targeted compounds, with promising results. This article critically reviews already published data on COX-2-selective inhibitors that have been tested in cancer clinical trials, highlights ongoing research and considers the future perspectives of this novel class of agents.