There have been tremendous advances in our understanding of cancer from the application of molecular biology over the past decade. The disease is caused by a series of defects in the genes that accelerate growth - oncogenes - and those that slow down cellular turnover - tumour suppressor genes. The proteins they encode provide a promising hunting ground in which to design and test new anticancer drugs. Several treatment strategies are now under clinical trial entailing direct gene transfer. These include the use of gene marking to detect minimal residual disease, the production of novel cancer vaccines by the insertion of genes which uncloak cancer cells so making them visible to the host's immune system, the isolation and coupling of cancer specific molecular switches upstream of drug activating genes, and the correction of aberrant oncogenes or tumour suppressor genes. The issues in these approaches are likely to have a profound impact on the management of cancer patients as we enter the next century.

The key problem in the effective treatment of patients with solid tumours is the similarity between tumour cells and normal cells. Local procedures such as surgery and radiotherapy may be effective, but only if the malignant cells are confined to the area treated. This is so in around one third of cancer patients. For most, some form of systemic selective therapy is required. Though many cytotoxic drugs are available, only a small proportion of patients are actually cured by them. The success stories in Hodgkin's disease, non-Hodgkin's lymphoma, childhood leukaemia, choriocarcinoma, and germ cell tumours have simply not materialised for the common cancers such as those of the lung, breast, or colon. And this is despite enormous efforts in new drug development, clinical trials of novel drug combinations, the addition of cytokines, high dose regimens, and even …