Criteria for establishing that a virus is oncogenic

To prove that a particular infectious agent causes a disease is much more difficult in human subjects than in other animals for both ethical and practical reasons. Where the disease is a malignant tumour with a long latent period the situation is even more difficult. For these reasons, it is often necessary to concentrate in the first instance on association of the virus with the disease, and this is discussed in the context of papillomaviruses. Association of a virus with a tumour may occur for a variety of reasons other than the virus being the cause of the tumour. This is illustrated by several examples of parvoviruses and DNA tumour viruses. Conversely, the absence of any sign of virus or viral nucleic acid in a tumour does not prove that the tumour was not induced by a virus. Apart from association of a virus with a tumour it is also necessary to show that the virus in question is oncogenic. Again this cannot normally be done directly, so that indirect evidence from animal experiments or from in vitro transformation is likely to be the best available alternative. In the final analysis the best proof of oncogenicity may be the effectiveness of intervention directed at the virus.