Premium
Neural network model of orbofiban pharmacodynamics from sparse phase‐II data
Author(s) -
Mager D. E.,
Shirey J. D.,
Cox D.,
Fitzgerald D. J.,
Abernethy D. R.
Publication year - 2005
Publication title -
clinical pharmacology and therapeutics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.941
H-Index - 188
eISSN - 1532-6535
pISSN - 0009-9236
DOI - 10.1016/j.clpt.2004.12.244
Subject(s) - pharmacodynamics , clinical pharmacology , medicine , population , drug , pharmacology , artificial neural network , pharmacokinetics , machine learning , computer science , environmental health
Background/Aims The purpose of this study was to develop a neural network (NN) pharmacodynamic (PD) model that correlates the inhibition of ex vivo platelet aggregation by orbofiban, an oral GPIIb/IIIa antagonist, with the administered dose and patient characteristics. Methods Data were obtained from a Phase‐II dose‐finding study in patients presenting with acute coronary syndromes. A back‐propagation NN was designed to predict drug effect measured at pre‐dose and 4 and 6 hours on treatment days 1, 28, and 84 (9 responses/patient). The training set (TS) consisted of patients for whom complete response profiles were reported (n=67), and remaining patients were included in the validation data set (VS; n=47). The concentration‐effect relationship was described also using a population inhibitory sigmoidal model, and a comparison of the predictive performances of both models was performed. Results The final NN reasonably described orbofiban PD from sparse data sets (r 2 =0.83 & 0.61; TS & VS) without specifying a structural model or drug concentrations. Despite considerable inter‐patient variability in response‐time profiles, the population model revealed a strong correlation between drug concentration and effect and exhibited greater precision than the NN model. Conclusions Although the population model showed greater precision, these results suggest that NNs may be useful for predicting drug PD when plasma concentrations are relatively unpredictable or unavailable. Clinical Pharmacology & Therapeutics (2005) 77 , P92–P92; doi: 10.1016/j.clpt.2004.12.244