Table of Contents Author Guidelines Submit a Manuscript
Canadian Journal of Infectious Diseases
Volume 5 (1994), Suppl A, Pages 51A-59A

Monte Carlo Comparison of Rival Experimental Designs For Two-Agent Combined Action Studies

William R Greco, David C Sutor, John C Parsons, Leonid A Khinkis, Lily Hsieh, Sowmya R Rao, Ying Tung, Christopher C Currie, and Robin Susice

Department of Biomathematics, Roswell Park Cancer lnstitute, New York State Department of Health, Buffalo, New York, USA

Copyright © 1994 Hindawi Publishing Corporation. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.


OBJECTIVE: The combined action of two or more chemotherapeutic agents and/or biological agents can be quantitatively described wilh empirical multidimensional concenlration-effect response surface models. This intuitive statistical approach provides a framework for suggesting experimental designs for in vitro. in vivo and possibly clinical experiments of agent combinations. Five rival 32-point experimental designs for in vitro continuous response two-agent combined action studies were compared using Monte Carlo simulation.

DESIGN: The designs were: factorial; central composite: one-ray in duplicate; four-ray; and D-optimal.

SETTING: Datasets were simulated by generating ideal data with the authors’ flagship two-agent combined action model. which includes six parameters: the control sunrival Econ=100 (where Econ is the full range of response that can be affected by the drug); median effective concentrations. IC50.1=10. IC50.2= 1 for drug 1 and drug 2, Respectively; slope parameters. m1 =- 1. m2=-2 for drug 1 and drug 2. respectively; and the interaction parameter, α=1 or α=5. For each design, for each of four types of error (absolute. relative with 1% coefficient of variation [cv]. relative with 10% cv. and relative with 10% cv plus a noise constant of 1% of Econ) . for each of two values of the true α (1, 5). 500 Monte Carlo datasets were generated. and then flt via weighted nonlinear regression wilh lhe flagship model.

MAIN Results: For the α parameter. for relative error-containing datasets. the D-optimal designs had the smallest variances.

CONCLUSION: The counterintuitive D-optimal designs may be useful for studies in which the experimental units are relatively precious. and frugal designs are essential. In addition. it may be fruitful to add the D-optimal design points lo standard experimental designs.