Design of trials for antibody treatment of tumour.

Our discussion is summarized in Fig. 1. Cytotoxic drugs are shown with the classical sigmoid relationship of response to log dose, with a parallel increase in toxicity which suits the Phase I approach. With antibody it is doubtful that clinical tumour ever responds to single doses with an asymptote of 100% ablation, so arbitrarily lower asymptotes have been used, higher for immunotoxins than for antibodies (n). The response curves have been made sigmoid simply by selecting, among the infinity of mathematical functions, two hypothetical functions of dose which yield sigmoid transformations of the data. After superimposing on the graphs the occurrence of toxicity, one discerns with antibodies (n) only a threshold dose, above which any toxicity will be highly dependent on the extent of encounter with readily accessible cellular or molecular antigen. Immunotoxins display an increasing toxicity with dose, but the positioning of the toxicity gradient in relation to dose is uncertain because of the ameliorating effect of uptake by tumour. The toxicity of an antibody and its derivatives can and should be clearly documented, but a case is presented here against trials of these reagents which systematically seek MTDs with only subsidiary attention to therapeutic effects. With concurrent evaluation of therapeutic effects. With concurrent evaluation of therapy and toxicity--in studies labelled Phase I/II in the current nomenclature--it may prove just as appropriate to relate toxicity to therapeutic efficacy as to dose.