Gene therapy for cancer has evolved over the last 20 years from its initial proof of principle strategies involving plasmid vectors and non-replicating genetically modified viruses to the current era to the extensive evaluation of live replicating oncolytic viruses (OV). Importantly clinical application has progressed from the phase I ‘proof of principle’ type studies to the randomised phase III clinical trial arena. Remarkably OV appear in the medical literature on a number of occasions over the last 100 years including prospective clinical studies in the 60’s and 70’s [1]. The finding from these studies have been regarded more as medical curiosities but evolving understanding of both molecular pathology of common malignancies as well as basic viral biology has bought the two fields back together again. Around 20 viruses have been or are been tested preclinically and a dozen or so viruses have been used in clinical trials over the last century. There are currently 7 viruses in clinical trial at present. Broadly there is a natural division between wild type and engineered viruses and for the engineered agents, a variety of transgenes ranging from immunostimulatory molecules through to tumour-associated antigens. However, opinion is still divided as to whether the anti-tumour efficacy of OV is due to direct cellular cytotoxicity or whether indeed oncolytic viruses are effectively a sophisticated immunotherapy where tumour cell lysis by OV primes an anti-tumour response which is then sustained by further rounds of oncolysis. The prototype OV were adenoviruses which evolved from their replication defective versions which in turn were problematic due to inadequate transduction efficiency. The ONYX 015 was the first administered replication competent engineered OV in humans and was evaluated in patients with malignancy in the pancreas, brain and head and neck, both as single agent and in combination with chemotherapy [2]. A similar agents has been licensed in China for the treatment of head and neck cancer. A number of other viruses are being tested through systemic and local delivery in humans. Reoviruses are naturally occurring and ubiquitous non-pathogenic virus. Cancer cells are susceptible to reovirus cytotoxicity partially as cancer cells are unable to activate PKR in response to viral double stranded RNA unlike normal cells. Clinical evaluation of reovirus alone or in combination with chemotherapeutics and radiation therapy is the largest OV programme currently ongoing worldwide. Evidence of anti-tumour activity has been documented radiologically and in tumour biopsies post treatment [3]. Four strains of engineered herpes simplex virus type 1 (HSV-1) have progressed to clinical assessment. These viruses have deletions in the gene including ICP 34.5 to support tumour selectivity and diminish neurovirulence. The most advanced programme is using the Oncovex agent; the ICP 34.5 gene replaced by a sequence encoding GMCSF to enhance anti tumour responses. The results of the current Oncovex programme are impressive with evidence of distance immune responses after local injection of patients of recurrent malignant melanoma. Newcastle Disease Virus is an apathogenic virus in humans and has been evaluated in over 100 patients with documented objective responses. Finally there has been a large clinical programme evaluating oncolytic vaccinia The Wyeth strain has been modified by deletion of thymidine kinase and incorporation of GM-CSF the recent reports are successful phase I study [4]. There are a number of issues that are currently challenging the development of OV therapy. A robust regulatory framework now exists in the UK so many of the problems are focused on practical issues such as handling of genetically modified agents within the Health Service, necessity for specialised pharmacy facilities, and defining where the agents are administered and any containment required. The other hurdles are factors limiting the anti-cancer efficacy. These include the access of the virus to tumours after injection which is limited by a profound and early neutralising immune response against the virus [5]. Other issues are how to measure the anti-tumour responses secondary to viral treatment. Convention uses traditional RECIST response criteria it may be less appropriate to these agents bearing in mind their tropism for tumour vasculature and propensity for tumours to necrose. Finally it is unlikely that OVs will have sufficient efficacy as single agents. Combination with cytotoxic and more targeted therapies, are the realistic route to adoption as a new anti-cancer modality. To date it is clear that current co-administration of cytotoxic agents does not damage the viruses and does not lead to enhanced toxicity to the patients.