OR WAIT 15 SECS
Even though the pharmaceutical industry has come to recognize the importance and associated benefits of using computers, evolving computer-based simulation technology has still to make a suitable impression.
Computers have revolutionized daily life and have become an indispensable technology solution. The pharmaceutical industry can be somewhat conservative in adopting new technologies, but most companies quickly realized that computers could potentially improve productivity and generate cost efficiencies by automating processes and speeding communication.
Computers playing a driving role in R&D is more recent and started in the area of drug design. Most people still associate computer technologies with drug discovery, but they are exerting a growing impact further up the R&D chain into clinical trials. With many companies expanding the application of technology solutions to a wider range of R&D areas, commercial departments are also using computer-generated information to help them in their R&D decision-making as these models draw together disparate pieces of information related to R&D, such as potential sales, technical risk, competitor activity, opportunity costs and strategic fit within the company's product portfolio. Several companies have developed their own customized models in-house that couple the best features of external models with their specific corporate needs. Financial and marketing models are also frequently used by companies at both early and late stages of drug development.
A development that should accelerate the adoption of computer-based simulations is the decision of regulatory agencies in Europe and the US to explore the use of these technologies for trials. If this technology can help enhance drug safety, it will provide an additional tool to help the agencies meet public expectations.
Within the discovery stage of R&D, computer technologies have found common use in areas such as pharmacophore identification, virtual screening, and quantitative structure activity relationship methods for lead optimization and absorption, distribution, metabolism, excretion and toxicity prediction.1 For example, computer simulations can be used to develop drugs based on the 3D-structures of the proteins targeted by the test compounds being developed. These systems can be used to screen the most promising compounds from those that are being developed. Advances in 3D computer simulation mean that the technology can be quickly used to predict the interaction between a protein and the candidate compound structures. Previously, more conventional simulations were often insufficient and unreliable in predicting such interactions and had proved of little value in the drug development process. To obtain the information they need, many companies have now turned to alternative simulation-based tests.
After 2000, following completion of the Human Genome Project, advanced computing solutions were sought by companies embarking on genomics programmes. Huge amounts of data are predicted to emerge from these programmes, and companies must determine how to process the vast array of genetic information and how to make decisions efficiently on what to move forward with and what to discard. In 2002, researchers from Curagen (CT, USA) calculated that existing drugs on the market only focused on 272 discrete molecular targets.2 In contrast, the amount of druggable targets, based on the human genome, was considered to be much higher at 8000. According to this analysis, 4990 were potential small-molecule targets, 2329 were antibody targets and 794 were targets for protein therapeutics.
Despite the opportunities offered by genomics, companies must ensure that they are set up to handle and manipulate the overwhelming amount of data; without an appropriate strategy to manage the IT approaches used, the availability of too many targets has the potential to actually slow the R&D process down and increase costs by congesting the new drug pipeline.3
Although discovery is an important part of the R&D process, clinical trials account for the majority of the R&D budget — and this share is increasing. Typically, pharma companies will devote approximately 40% of their R&D budget to clinical trials. Given the changing demands of regulators, this spend may need to be increased to run additional mandated trials. As a result, the failure of a drug at the later stages of clinical development is an extremely expensive failure for a company to bear.
Many of the compounds in clinical development fail because of problems regarding absorption, distribution, metabolism, elimination or toxicity of the drug in the body; therefore, much more effort is being put into eliminating these compounds before they enter the costly clinical trials process. Traditionally, this has been done through in vitro tests, but many companies are beginning to use computer simulations of human biological systems. These computer models, so-called in silico technology, allow researchers to set up different scenarios for the drug development process and the compounds progressing through them. There is hope that the in silico approach, used alongside other more traditional approaches, can help researchers choose the drug candidates that have the greatest chance of succeeding through clinical development. However, the beauty of this approach is that it can also be used both to kill off compounds that have a poor chance of making it through and to save compounds that were incorrectly deemed to have failed.
The author says...
These models can also be used to improve the trial itself. For example, during early-stage trials, simulations might focus on dose selection and the design of the study. Later on, the technology can be used to understand more about the study population; particularly to distinguish between potential responders and nonresponders to a drug. This information can then be used to reassess the study inclusion and exclusion criteria. Another valuable angle is that simulations can identify patients who may experience adverse events. The idea of simulations is to create more certainty among the clinical trial parameters to enhance the chances of success.
A number of companies believe they can save pharma companies millions of euros and are marketing software that can be applied to the drug development process. In 2005, PricewaterhouseCoopers predicted that in silico methods to create virtual trials might save 10% of resources in the short term, with greater savings to follow in the longer term.4 If the benefits of such technologies could be demonstrated, they would be very attractive to the pharma industry. So far, the signs have been extremely promising and a number of the major companies have set up dedicated simulation groups at their R&D centers.
One of the leaders in the in silico trial field is PharSight (CA USA), whose technology looks at how variability amongst patients might affect the results of a proposed clinical trial. Using the PharSight Trial Designer, researchers examine the different parameters for the clinical trial, and mathematical modelling tests the impact of these assumptions. The company's technology has been used by a number of the top pharmaceutical companies, including Pfizer, sanofi aventis and AstraZeneca. In one well-publicized case, Aventis used Pharsight technology to support a decision against taking a new selective oestrogen receptor modulator drug into clinical development.5 The savings from this decision were estimated at $50 million–$100 million. Entelos (CA, USA) is another player whose technology has been used by major companies; its biosimulation system was used by Johnson & Johnson for a Phase I diabetes trial, and showed how a 40% reduction in time and a 66% reduction in the number of patients could be achieved.6
Interestingly, the activity of these companies has attracted attention from regulatory agencies. In 2001, Pharsight signed a 3-year Cooperative Research and Development Agreement with the FDA's Center for Drug Evaluation and Research to focus on industry and regulatory needs for population pharmacokinetic/pharmacodynamic (PK/PD) modelling and clinical trial simulation.7 This work forms part of the FDA's Critical Path Initiative (CPI), which uses current thinking to develop new ways to improve the drug development process. The CPI has already identified bioinformation sciences as a key knowledge area to be leveraged. In Europe, the EMEA is also showing an interest in this field and held a conference in 2008 on applying computer modelling to paediatric drug development.8,9 This is because regulations require a specific evaluation of products in children because of the PK and PD differences between children and adults.
One of the problems identified in the slow uptake of in silico technology and other IT solutions among some companies is the reluctance of senior managers to relinquish authority over the decision-making process when compounds progress to clinical development. Despite advances in technology, it is still felt that human reasoning should take precedence because an in silico system could still result in errors; even though simulation models can factor in items such as the cost and success rates of clinical trials, and the potential market share and return on total investment generated by a new drug, some companies only see it as having a supporting role in drug development decision-making. Another factor is the shortage of people with skills in this field.
However, there is no doubt that in silico approaches can demonstrate value if applied correctly and the growing interest of regulatory agencies may persuade companies to take a fresh look at this technology area.
Faiz Kermani is a freelance consultant and President of the Global Health Education Foundation, a charity that supports medical education and medical research projects in developing countries. He is a member of Pharmaceutical Technology Europe's Editorial Advisory Board. email@example.com
1. H. Hong et al., SAR and QSAR in Environmental Research, 16(4), 339–347 (2005).
2. K. Davies, Bio IT World (2002). www.bio-itworld.com
3. G.M. Hampton and K. Sikora, Genomics in Cancer Drug Discovery and Development, Volume 96 (Advances in Cancer Research) (Academic Press, Amsterdam, 2007).
4. Pharma 2005: An Industrial Revolution in R&D (PricewaterhouseCoopers, 2005).
5. G. Shaw, Drug Discov. Dev., Quarter 1(3) (2006). www.dddmag.com
6. Entelos, Inc., Use of Systems Biology in Clinical Development: Design and Prediction of a Type 2 Diabetes Clinical Trial (2004). www.entelos.com
7. Pharsight, "FDA-Pharsight CRADA" (2009). www.pharsight.com
8. EMEA, EMEA workshop on modeling in paediatric medicines (14–15 April, London, UK, 2008). www.emea.europa.eu
9. E. Manolis and G. Pons G, Br. J. Clin. Pharmacol., Accepted Article (2009).