Biomedical Engineering Reference
In-Depth Information
Alternatively, the drug could be synthesized for the patient's specific genetic markers—as in tumor-
specific chemotherapy, for example. This synthesized drug might take a day or two to develop, unlike
the virtually instantaneous drug cocktail, which could be formulated by the corner pharmacist. The
tradeoff is that the drug would be tailored to the patient's genetic profile and condition, resulting in
maximum response to the drug, with few or no side effects.
How will this or any other killer app be realized? The answer lies in addressing the molecular biology,
computational, and practical business aspects of proposed developments such as custom
medications. For example, because of the relatively high cost of a designer drug, the effort will
initially be limited to drugs for conditions in which traditional medicines are prohibitively expensive.
Consider the technical challenges that need to be successfully overcome to develop a just-in-time
designer drug system. A practical system would include:
High throughput screening
— The use of affordable, computer-enabled microarray
technology to determine the patient's genetic profile. The issue here is affordability, in that
microarrays costs tens of thousands of dollars.
l
Medically relevant information gathering
— Databases on gene expression, medical
relevance of signs and symptoms, optimum therapy for given diseases, and references for
the patient and clinician must be readily available. The goal is to be able to quickly and
automatically match a patient's genetic profile, predisposition for specific diseases, and
current condition with the efficacy and potential side effects of specific drug-therapy options.
l
Custom drug synthesis
— The just-in-time synthesis of patient-specific drugs, based on the
patient's medical condition and genetic profile, presents major technical as well as political,
social, and legal hurdles. For example, for just-in-time synthesis to be accepted by the FDA,
the pharmaceutical industry must demonstrate that custom drugs can skip the clinical-trials
gauntlet before approval.
l
Achieving this killer app in biotech is highly dependent on computer technology, especially in the use
of computers to speed the process testing-analysis-drug synthesis cycle, where time really is money.
For example, consider that for every 5,000 compounds evaluated annually by the U.S.
pharmaceutical R&D laboratories, 5 make it to human testing, and only 1 of the compounds makes it
to market. In addition, the average time to market for a drug is over 12 years, including several
years of pre-clinical trials followed by a 4-phase clinical trial. These clinical trials progress from safety
and dosage studies in Phase I, to effectiveness and side effects in Phase II, to long-term surveillance
in Phase IV, with each phase typically lasting several years.
What's more, because pharmaceutical companies are granted a limited period of exclusivity by the
patent process, there is enormous pressure to get drugs to market as soon as a patent is granted.
The industry figure for lost revenue on a drug because of extended clinical trials is over $500,000 per
day. In addition, the pharmaco-economic reality is that fewer drugs are in the pipeline, despite
escalating R&D costs, which topped $30 billion in 2001.
Most pharmaceutical companies view computerization as the solution to creating smaller runs of
drugs focused on custom production. Obvious computing applications range from predicting efficacy
and side effects of drugs based on genome analysis, to visualizing protein structures to better
understand and predict the efficacy of specific drugs, to illustrating the relative efficacy of competing
drugs in terms of quality of life and cost, based on the Markov simulation of likely outcomes during
Phase IV clinical trials.
Despite these obvious uses for computer methods in enabling the drug discovery and synthesis
process, the current state of the art in these areas is limited by the underlying information
technology infrastructure. For example, even though there are dozens of national and private
genome databases, most aren't integrated with each other. Drug discovery methods are currently
limited to animal and cell models. One goal of computerizing the overall drug discovery process is to
create a drug discovery model through sequencing or microarray technology. The computer model
would allow researchers to determine if a drug will work before it's tried on patients, potentially
bypassing the years and tens of millions of dollars typically invested in Phases I and II of clinical
trials.
