Cancer Prevention Epidemiology and Biostatistics
Senior Faculty, Director of Biomathematics & Biostatistics
Roswell Park Cancer Institute
Elm and Carlton Streets
Buffalo New York USA 14263
Tel: 716-845-8641
Fax: 716-845-8467
E-mail: william.greco@roswellpark.org

Program: Pharmacometrics; Agent Combinations
The Pharmacometrics Unit integrates the main components of the modeling process, the derivation of models, experimental design, the fitting of models to data and the simulation of data from models, into the laboratory programs of other scientists to enable the research group to focus more deeply into complex systems, such as those involving metabolic biochemical pathways, receptors, enzymes, cell growth, cell cycle kinetics, tumor growth, cell-cell interactions and pharmackinetics.

Progress
A project was initiated to study and mathematically model the effect of chemotherapeutic agents from several classes on gene expression from chemosensitive and resistant tumor cells (in vitro cell culture) with Affymetrix GeneChip oligonucleotide arrays, and with spotted cDNA microarrays. The project will leverage the concentration-time-effect modeling already published by Dr. Greco's group, which focused on a much simpler endpoint, cell proliferation, and the experience of 4 laboratories with specific anticancer agents and specific human cell lines. Each laboratory will study a single anticancer agent, with a single common human cell line, with the addition of a resistant subline for 2 of the projects.

The overall hypotheses are: (a) the exposure of cancer cells to anticancer agents will result in patterns of change in time and agent concentration of gene expression; (b) These patterns will be amenable to mathematical modeling; (c) This mathematical modeling will be useful for elucidating mechanisms of sensitivity and resistance of cancer cells to anticancer agents.

Currently there are no published studies of the effects of anticancer agents on gene expression. The designed redundancy among laboratories should lead to a definitive assessment.

General Research Interest
The Biomathematics/Biostatics unit assists researchers in the conduct of the scientific method, which includes derivation of mathematical/statistical models, experimental design, gathering and management of data, fitting of models to data, and simulation of data from models. Concepts and tools from the disciplines of statistics, engineering science and computer science are used to accomplish these goals. Most projects require the use of standard math/stat tools, which are accessd via problem requires an innovative approch, this group develops novel biomathematical/biostatistical concepts and methodologies to solve these problems.
Research collaborations of the Biomathematics/Biostatistics unit also have an educational component. The faculty provides educational support to RPCI researchers through one-on-one tutorials, seminars, workshops, formal classes and research experiences for apprentice quantitative scientists.

Students

The overall research focus recognizes the ongoing paradigm progression in Modern Biology from Biochemistry to Molecular Biology to Quantitative Biology. A contribution to this 3rd wave of Modern Biology is a method that was created, developed and tested for determining whether the combination of two agents is synergistic, additive or antagonistic. The method consists of fitting a set of equations in a hierarchical manner to dose-response data with nonlinear regression or maximum likelihood estimation, choosing the model which best fits the data, and interpreting the resulting parameter estimates. The new method overcomes many of the problems associated with older traditional methods. The new method has been applied extensively to both simulated and real data. Research is ongoing to create new mathematical/statistical models useful for describing complex concentration-time-effect phenomena in Cancer Pharmacology. These semi-empirical models are used for developing leads to the mechanistic understanding of drug resistance, selective toxicity and response heterogeneity.

Key Publications

 

• Bucki C, Khinkis L, Malmberg A, Parsons J, Greco WR. Monte Carlo investigation of rival experimental designs for the Hill model. In Monte Carlo Simulation: Proceedings of the International Conference on Monte Carlo Simulation, Monte Carlo, Principality of Monaco, 18-21, June, 2000. (G.I. Schueller & P.D. Panos, eds), A.A. Balkema Publishers, Lisse, The Netherlands 2001, pp. 247-253.
• Greco WR. Commentary on Borgert et al. (2001) Evaluating chemical interaction studies for mixture risk assessment - Comments. Human and Ecological Risk Assessment 7:306, 2001.
• Greco WR. Review of Conference, Critical Assessment of Techniques for Microarray Data Analysis, CAMDA’00, Duke University, 12/18-12/19, 2000, Special Interest Group on Biomedical Computing of the Association of Computing Machinery, pp. 26, 2001.