Introduction

While approximately half of all patients with cancer can be cured, either by locoregional therapy (surgery or radiation) or by systemic chemotherapy, this success is not gained without a price – that of toxicity to normal tissues, either during treatment or afterwards. Prominent among the late effects of treatment is the development of second malignancies, for radiation and chemotherapy are both carcinogenic (Boice, 1988; Boffetta and Kaldor, 1994). The protean toxic manifestations of cancer therapy arise because of the lack of selectivity of present treatment approaches, a shortcoming that is hardly surprising when it is considered that the therapeutic effects of radiation and chemotherapy were first recognized, and have subsequently been enhanced, almost entirely as a consequence of empirical observations. Thus, all new approaches to cancer treatment are directed towards improving the therapeutic index, i.e., decreasing toxicity while achieving similar, or preferably better, tumor cell kill. While empirical clinical trials remain the mainstay of research directed towards the improvement of therapeutic results, it seems probable that optimal therapeutic indices, in which curative therapy is associated with minimal toxicity (i.e., tumor-specific therapy), will only be achieved by the identification of major biological differences between neoplastic cells and their normal counterparts that can be exploited therapeutically (i.e., the identification of an ‘Achilles heel’ in the tumor cell).