Interactive technique for optimizing drug development from the pre-clinical phases

Medicilon’s Chemistry department provides custom synthesis, medicinal chemistry, focus library preparation, and process and scale-up synthesis. Either in FTE or integrated project team arrangement, our medicinal chemistry team has helped clients discover novel, potent, and highly druggable chemical series. Our contribution has led to several preclinical and clinical candidates. We also provide scale up service to deliver kilogram of material with high-quality and on-time delivery.

Contact us:marketing@medicilon.com.cn web:www.medicilon.com

A method of performing interactive clinical trials for testing a new drug. A pre-clinical phase is performed in which a computer model for pharmacokinetics and pharmacodynamics of the drug is created and adjusted based on in vitro studies and in vivo studies in animals. A phase I clinical research is performed in which a clinical trial on at least a single dose is performed in parallel with performing computer simulation studies using the computer model. An optimal protocol is determined for the most responsive patient populations and indications for a phase II clinical trial. Phase II clinical trial is performed where a number of small scale clinical trials are performed in parallel based on results of the above. Phase III clinical research is performed for chosen indications by chosen protocols. Phase IV studies are performed for post-marketing subpopulation analysis and long term product safety assessment.

The drug industry is facing substantial challenges with regards to cost-containment and time-to-market for its high-potential candidates. Currently pharmaceutical companies investigate many different methods for increasing their productivity in the development process in order to compensate for increasing difficulties in recouping the investment in drug development.

However, the classical method of clinical trials design suffers major drawbacks. On the one hand, developing drugs by “trial and error” alone can not guarantee that the selected schedules are better than other, yet to be tried, treatment regimens. On the other hand, the number of schedules which can be empirically tested is negligibly small with respect to the potential number of sensible schedules.

Research shows that the effects of the drug may crucially correlate with the internal dynamics of the tumor growth processes, as well as with the relevant patient's physiology. These aspects might often be too complex to be estimated by the naked eye, and slight nuances in the treatment schedule may be critical for the effect achieved. In theory, if all potential treatment schedules could be tested, considering all the available information on the involved biological processes, pathological processes and the momentary effect of the drug on every element of these processes, one could, a-priori, suggest a theoretical set of the most promising treatment schedules for a given indication, or, even, for a given patient. Subsequently, these promising schedules would be clinically tested, thus saving human resources and time, and helping to achieve maximal possible therapeutic effects of the tested drug.

Needless to say that such methods would enable to rehabilitate drugs with valid properties, which failed during the development process, due to insufficient efficacy, or limitations of toxicity, which could possibly be overcome by modifying the treatment schedule. In addition, these methods would enable a “Go-NoGo” decision to be made early during the clinical trial process.

Today, there exist elaborate and highly interdisciplinary and multidisciplinary methods, which can employ modern computing facilities for integrating the enormous body of relevant biological, medical, pharmacological and mathematical (dynamical) information into comprehensive systems for simulating different drug treatment scenarios. The techniques disclosed herein are based on more than two decades of biomathematical research in the area of disease control optimization. Thus, mathematical algorithms have been developed, which simulate the dynamics of key biological, pathological and pharmacological processes in a patient undergoing drug treatment, either by monotherapy, or by combination of cytotoxic and/or cytostatic agents, and/or by growth-factors. This set of computerized mathematical models, in conjunction with advanced optimization algorithms have now yielded an in silico patient engine, having a range of applications designed to deliver optimal drug treatments for cancer and hematological disorders.

Disclosed herein are techniques for improving anticancer drug development, which employ such an in silico patient engine in drug development. The disclosed techniques enable the drug developer an ongoing dialogue, from pre-clinical phase through Phase-IV, for generating, fine-tuning and validating a reliable drug/disease/host model. Thus, relatively early during development, i.e., by the end of Phase-I, and no later than in mid-Phase-II, the model already contains the precise PK/PD drug parameters, to be implemented in the in silico patient simulations. At this stage numerous drug schedules (termed “infinite protocol space”) are simulated for any desired indication, and proprietary optimization techniques are employed for selecting, among the vast number of simulation scenarios, those yielding best results according to the list of specifications set by the drug developer. In this way one identifies the most appropriate indications/monotherapy/combination treatments for the drug. At this early stage a “Go-NoGo” decision can be made.

Following the disclosed techniques clinical trials can be rationally designed, which will be based upon a gradual improvement and zeroing-in on the best prediction-directed treatment schedules. It is important to stress that the disclosed technique carries little risk of yielding false predictions, since the algorithm has been designed so as to be continuously validated and improved by information derived in parallel from clinical trials.