Optimizing the use of ultrasound to deliver chemotherapeutic agents to cancer cells from polymeric micelles

In this study, we present an artificial neural network (ANN) model that attempts to predict the dynamic release of doxorubicin (Dox) from P105 micelles under different ultrasonic power densities at 20 kHz. The goal is to utilize the developed ANN model in optimizing the ultrasound application to achieve a target drug release at the tumor site by controlling power density and ultrasound duration via an ANN-based model predictive control. The parameters of the controller are then tuned to achieve good reference signal tracking. (C) 2010 The Franklin Institute. Published by Elsevier Ltd. All rights reserved.