Related Experiment Videos
Mapping the dose-effect relationship of orbofiban from sparse data with an artificial neural network
Donald E Mager1, Jason D Shirey, Dermot Cox
1Laboratory of Clinical Investigation, National Institute on Aging, Gerontology Research Center, Baltimore, Maryland, USA.
Journal of Pharmaceutical Sciences
|October 4, 2005
Summary
A neural network model showed promise for predicting drug effects, but a population model demonstrated greater precision in correlating drug concentration with patient response. Neural networks may aid personalized medicine when drug levels are uncertain.
Area of Science:
- Pharmacology
- Computational Biology
- Clinical Trials
Background:
- Glycoprotein IIb/IIIa antagonists are crucial in managing acute coronary syndromes.
- Understanding drug pharmacodynamics is essential for optimizing therapeutic outcomes.
- Orbofiban is an oral glycoprotein IIb/IIIa antagonist investigated for its efficacy.
Purpose of the Study:
- To develop and evaluate a neural network (NN) model for predicting orbofiban's pharmacodynamic effects.
- To compare the predictive performance of the NN model with a population direct-effect inhibitory sigmoidal model.
- To explore the utility of NNs in individualizing pharmacotherapy.
Main Methods:
- A back-propagation neural network was designed to correlate orbofiban dose and patient characteristics with ex vivo platelet aggregation inhibition.
- Data from a Phase-II dose-finding study in acute coronary syndrome patients were utilized.
- The NN model's predictive performance was compared against a population sigmoidal model.
Main Results:
- The neural network model reasonably described orbofiban pharmacodynamics using sparse data.
- The population model identified a strong correlation between drug concentration and effect, showing greater precision than the NN model.
- Significant inter-patient variability in response-time profiles was observed.
Conclusions:
- While the population model offered higher precision, neural networks show potential for personalizing pharmacotherapy.
- NNs may be particularly useful when drug concentrations are unpredictable or unavailable.
- Further research into NN applications in pharmacotherapy is warranted.