Monoclonal antibodies to tumour-associated antigens have great theoretical potential for the specific targeting of radioactivity and anti-neoplastic agents to tumours. The clinical success of monoclonal antibody-based cancer diagnosis and therapy depends, however, on solving a number of pharmacokinetic delivery problems. These include: (i) slow elimination of monoclonal antibodies from the blood and poor vascular permeability; (ii) low and heterogeneous tumour uptake; (iii) cross-reactivity with normal tissues; (iv) metabolism of monoclonal antibody conjugates; and (v) immunogenicity of murine forms in humans. As a result of extensive pharmaceutical and pharmacokinetic research conducted over the past 10 to 15 years, several potential solutions to these delivery problems have been identified. Blood concentrations of antibody conjugates may be reduced through regional administration, the use of antibody fragments, interventional strategies and various pre-targeting techniques. Tumour uptake may be increased through administration of higher doses, or the use of agents to increase tumour vascular permeability. Tumour retention of antibody conjugates may be improved by inhibition of metabolism, by using more stable linkage chemistry. Alternatively, normal tissue retention may be decreased through the use of metabolisable chemical linkages inserted between the antibody and conjugated moiety. Very small antigen-binding fragments and peptides that exhibit improved tumour penetration and more rapid elimination from the blood and normal tissues have been prepared by genetic engineering techniques. Chimeric (mouse/human) and human monoclonal antibodies have been developed to circumvent the problem of immunogenicity. Future research will continue to be focused on improvements in the design of monoclonal antibodies for tumour targeting, with the ultimate goal of finally uncovering the 'magic bullet' envisioned by Paul Ehrlich almost a century ago.