Carboplatin is a potent anticancer agent that has shown efficacy in clinical trials against malignant glioma, one of the most deadly cancers in humans. However, a high systemic dose is required to achieve an effective concentration in the brain because of the presence of the blood-brain barrier (BBB). Such a high dose can cause many side effects. Local delivery of antitumor agents to the brain using injectable and biodegradable microspheres is a new strategy for the treatment of malignant glioma. This method is able to bypass the BBB and allows maximal local exposure and minimal systemic exposure to avoid the severe side effects of carboplatin. Delivering sustained-release microspheres directly to the tumor site could also control local tumor recurrence and improve survival. In the present studies, carboplatin-loaded microspheres were delivered intracerebrally in rats. No signs of systemic or neurologic toxicity associated with the microspheres implanted in the rat brain were observed. The in vivo release of carboplatin followed apparently zero-order release kinetics up to 30 days. The surface characteristics of the microspheres retrieved from the rat brains changed with the progress of polymer biodegradation. Implantation of the microspheres evoked a transient and localized inflammatory reaction that was well tolerated by the animals.

The purpose of this work was to develop a computer program that assists optimization of controlled-release devices, both visually and mathematically, using response surface methodology (RSM). A Windows-based computer program, Optima, which interactively implemented a number of subroutines for the optimization procedure, was developed. Optima is an integrated, user-friendly, and graphically oriented program for pharmaceutical dosage form optimization. Central composite design is implemented in the program. First- and second-order models containing up to five variables can be fitted to the data. The user can also choose between linear and exponential individual desirability functions, and use them to construct an overall desirability function that combines all the response variables in a single response. The program can predict the optimum levels of experimental variables, with respect to individual responses and/or the overall desirability. Optima has been successfully used in the development of sustained-release AZT-loaded microspheres. During the optimization process, three experimental variables were investigated and four responses were measured. The experimental design was a central composite design that was generated by the program. The response values were used by the program to calculate the individual desirability functions, which were then combined into an overall desirability function. The individual responses as well as the overall desirability function were optimized by fitting to a second-order polynomial equation. The response surfaces were generated and optimum levels of the experimental variables were predicted. The observed responses of the optimized formulation were very close to those predicted by Optima. The program proved to be a very useful, integrated tool for optimization of the controlled-release microspheres.